Search

This research brings a novel approach for current measurement using magnetic shape memory (MSM) smart alloys. The non-conventional instrument transformer (NCIT) proposed in this research uses the property of these alloys that their shape changes when exposed to a magnetic field. It has been shown that it is possible to measure alternating currents (a.c.) in high voltage overhead transmission lines by correlating the magnetic field produced by the current to shape changes in an MSM-based sensor. Methodologies for finite element modelling of the proposed NCIT have been developed. The developed methodology and obtained results are validated by comparing them to the results obtained through an experiment done by a manufacturer of MSM materials. 5M Ni-Mn-Ga MSM crystals with Type I twin boundaries and a load of 0.5 N/mm2 were identified as the most suitable type of MSM materials for this application. The combination of a very long fatigue life, with relatively low twinning stress, makes them the most prospective for use in MSM-based current sensors. The main characteristics of overhead transmission lines are described as well as the types of conductors typically used. This analysis brought us to the conclusion that special attention in this research should be given to ACSR and AAAC conductors, more specifically to 528-Al1/69- ST1A conductor (old code MOOSE) and 996-AL5 (old code REDWOOD). Additionally, the latest trends in the development of overhead transmission lines are discussed, as well as international standards which are relevant to these types of lines. These conductors were modelled in finite element (FE) package ANSYS APDL, together with the MSM element and the magnetic circuit, and included into a single finite element model. This approach allows us to take into account significant changes that take place within an MSM element during its elongation. Based on this, we were able to determine both the bottom and upper limits of the measurement range, optimise the NCIT for transmission lines, and propose several designs of the NCIT. Finally, this allowed relating the current inside the conductor to the voltage at the output of the LVDT.

Partially vegetated flows like those over rivers with riparian vegetation are common in natural environments. In these flows, vegetation acts as a porous obstruction for the flow. Thus, the slow interstitial fluid inside the vegetation and the rapid free streams above and next to it, lead to large gradients of shear at the vegetation patch boundaries, where shear layers are created. These shear layers are characterised by large turbulent coherent structures that interact with secondary circulations developed due to the differential roughness between the vegetation and the adjacent unobstructed flow. This hydrodynamic interaction dictates the interfacial mass and momentum transport in aquatic environments. As vegetation provides important ecosystem services, such as water quality improvement through nutrients and heavy metals retention, sediment trapping, and oxygen flux, the hydrodynamic transport occurring at the interface can significantly impact many physical, chemical and ecological processes. Therefore, flow characterisation through patchy vegetation is important for a system-level understanding of vegetated flows and the management of river and coastal systems. However, the characterisation of these flows is complex, owing to their three-dimensionality, which is not yet fully understood. This thesis presents the characterisation of the three-dimensional flow developed on a partially vegetated channel, focussing upon the role of secondary circulation on mass and momentum transport. Firstly, an experimental approach was employed to determine the flow properties, taking 3D-point velocity measurements, using an Acoustic Doppler Velocimeter (ADV), in the cross-sectional plane on a partially vegetated flume. Experiments were configured to examine the influence of vegetation density, over a range of reported realistic field conditions, and shallowness on the mean and turbulent flow. Subsequently, an analytical model was used to predict depth-averaged streamwise velocities, modelling the effects of secondary circulation on the momentum equation. The results reveal that secondary circulation cells scale with the vertical shear layer thickness within the vegetation and the water depth beside the canopy patch. The rotational direction of the major pair of cells located at the lateral canopy-water interface inverts, either at high densities due to the growth of an emergent rotating pair of cells, or at reduced water depths. Secondary circulation contributions to vertical mass transport, for the higher submergence condition, are estimated to be of the same order of magnitude as that of turbulent diffusion. Advective dispersive stresses arising due to mean flow fluctuations are found to be up to 21% of the total vertical shear stress and the main contributors to lateral shear stress within the vegetation and in the main channel outside the mixing layer. Additionally, this work presents the first attempt at using the Shiono and Knight Method, an analytical model based on the depth-averaged Navier-Stokes equation, to model depth-averaged streamwise velocities in submerged partially vegetated channels. Appropriate advice on the values of calibration parameters and cross-section discretisation is given. The combined experimental and modelling approaches provide and insight into the 3D flow structure and highlights the importance of secondary circulation when evaluating mass and momentum transport in submerged partially vegetated flows.

In railways, turnouts, or also called switches and crossings, are widely reported as the origin of many problems owing to their involvement in many accidents and the major share on the budget of infrastructure managers. Turnouts have an asymmetrical and large geometry in order to enable diversion between railway tracks, leading to complex dynamic behaviour of the structure. The complex structure of turnouts and its dynamic effect catalyse the process of many failure types. One of common failures over a turnout structure is dynamic fragility. Particularly, high-frequency high-magnitude impact forces that are observed over turnout crossing noses cause many dynamic fragility problems including long terms defects such as fatigue. This thesis aims at establishing a new numerical method to simulate dynamic fragility of turnout crossings from fracture mechanics' point of view. In order to achieve this aim, first, it is targeted to improve the understanding of dynamic behaviours of turnouts under impact loadings varying from ideal environments to extreme events. A new numerical methodology is proposed to evaluate the complex behaviours of turnouts and its components. The new insights stemmed from this research has shown the importance of taking dynamic effects into account and the force spectra observed over the turnout structure. More importantly, the Multiphysics informed models helped to determine the critical section over a turnout structure and corresponding failure modes. Following having a more comprehensive understanding of dynamic behaviours of railway turnouts, the fatigue behaviour of the critical section has been examined via Peridynamics. Peridynamics is a recent and promising numerical tool in the field of Fracture Mechanics. Prior to fatigue investigation, two new methodologies have been introduced into Peridynamics theory. Those methodologies have helped to track the formation of cracks, and reduce the simulation time significantly. It would not be possible to conduct fatigue simulations of complex geometries without those contributions. As is the final target, fatigue simulations have been carried over critical sections of a turnout as well as experimental geometries. It was found that the fatigue model in Peridynamics is promising in certain damage modes and confirms the general knowledge in most cases. Nonetheless, its efficiency has been found questionable in the case of fatigue over the crossing nose under a mixed mode compression loadings. The characteristics of Peridynamics have imposed certain limitations and it seems that those challenges must be addressed first in order to be able to increase the efficiency and to apply in railway applications.

Upper-limb prostheses are either too expensive for many consumers or exhibit a greatly simplified choice of actions, this research aims to enable an improvement in the quality of life for recipients of these devices. Previous attempts at determining the hand shapes performed during activities of daily living (ADL) provide a limited range of tasks studied and data recorded. To avoid these limitations, motion capture systems and machine learning techniques have been utilised throughout this study. A portable motion capture system created, utilising a Leap Motion controller (LMC), has captured natural hand motions during modern ADL. Furthering the use of these data, a method applying optimisation techniques alongside a musculoskeletal model of the hand is proposed for predicting muscle excitations from kinematic data. The LMC was also employed in a device (AirGo) created to measure joint angles, aiming to provide an improvement to joint angle measurements in hand clinics. Hand movements for 22 participants were recorded during ADL over 111 hours and 20 minutes - providing a taxonomy of 40 and 24 hand shapes for the left and right hands, respectively. The predicted muscle excitations produced joint angles with an average correlation of 0.58 to those of the desired hand shapes. AirGo has been successfully employed within a hand therapy clinic to measure digit angles of 11 patients. A taxonomy of the hand shapes used in modern ADL is presented, highlighting the hand shapes currently more appropriate to consider during upper-limb prostheses development. A method for predicting the muscle excitations of the hand from kinematic data is introduced, implemented with data collected during ADL. AirGo offered improved repeatability over traditional devices used for such measurements with greater ease of use.

Camera based methods for optical coordinate metrology are growing in popularity due to their non-contact probing technique, fast data acquisition time, high point density and high surface coverage. However, these optical approaches are often highly user dependent, have high dependence on accurate system characterisation, and can be slow in processing the raw data acquired during measurement. Machine learning approaches have the potential to remedy the shortcomings of such optical coordinate measurement systems. The aim of this thesis is to remove dependence on the user entirely by enabling full automation and optimisation of optical coordinate measurements for the first time. A novel software pipeline is proposed, built, and evaluated which will enable automated and optimised measurements to be conducted. No such automated and optimised system for performing optical coordinate measurements currently exists. The pipeline can be roughly summarised as follows: intelligent characterisation → view planning → object pose estimation → automated data acquisition → optimised reconstruction. Several novel methods were developed in order to enable the embodiment of this pipeline. Chapter 4 presents an intelligent camera characterisation (the process of determining a mathematical model of the optical system) is performed using a hybrid approach wherein an EfficientNet convolutional neural network provides sub-pixel corrections to feature locations provided by the popular OpenCV library. The proposed characterisation scheme is shown to robustly refine the characterisation result as quantified by a 50 % reduction in the mean residual magnitude. The camera characterisation is performed before measurements are performed and the results are fed as an input to the pipeline. Chapter 5 presents a novel genetic optimisation approach is presented to create an imaging strategy, ie. the positions from which data should be captured relative to part's specific geometry. This approach exploits the computer aided design (CAD) data of a given part, ensuring any measurement is optimal given a specific target geometry. This view planning approach is shown to give reconstructions with closer agreement to tactile coordinate measurement machine (CMM) results from 18 images compared to unoptimised measurements using 60 images. This view planning algorithm assumes the part is perfectly placed in the centre of the measurement volume so is first adjusted for an arbitrary placement of the part before being used for data acquistion. Chapter 6 presents a generative model for the creation of surface texture data is presented, allowing the generation of synthetic butt realistic datasets for the training of statistical models. The surface texture generated by the proposed model is shown to be quantitatively representative of real focus variation microscope measurements. The model developed in this chapter is used to produce large synthetic but realistic datasets for the training of further statistical models. Chapter 7 presents an autonomous background removal approach is proposed which removes superfluous data from images captured during a measurement. Using images processed by this algorithm to reconstruct a 3D measurement of an object is shown to be effective in reducing data processing times and improving measurement results. Use the proposed background removal on images before reconstruction are shown to benefit from up to a 41 % reduction in data processing times, a reduction in superfluous background points of up to 98 %, an increase in point density on the object surface of up to 10 %, and an improved agreement with CMM as measured by both a reduction in outliers and reduction in the standard deviation of point to mesh distances of up to 51 microns. The background removal algorithm is used to both improve the final reconstruction and within stereo pose estimation. Finally, in Chapter 8, two methods (one monocular and one stereo) for establishing the initial pose of the part to be measured relative to the measurement volume are presented. This is an important step to enabling automation as it allows the user to place the object at an arbitrary location in the measurement volume and for the pipeline to adjust the imaging strategy to account for this placement, enabling the optimised view plan to be carried out without the need for special part fixturing. It is shown that the monocular method can locate a part to within an average of 13 mm and the stereo method can locate apart to within an average of 0.44 mm as evaluated on 240 test images. Pose estimation is used to provide a correction to the view plan for an arbitrary part placement without the need for specialised fixturing or fiducial marking. This pipeline enables an inexperienced user to place a part anywhere in the measurement volume of a system and, from the part's associated CAD data, the system will perform an optimal measurement without the need for any user input. Each new method which was developed as part of this pipeline has been validated against real experimental data from current measurement systems and shown to be effective. In future work given in Section 9.1, a possible hardware integration of the methods developed in this thesis is presented. Although the creation of this hardware is beyond the scope of this thesis.

Fringe projection systems can acquire a point-cloud of more than a million points in minutes while not needing to ever physically touch the measurement surface and can be assembled using relatively inexpensive off-the-shelf components. Fringe projection system can conduct measurements faster than their tactile counterparts and typically require less training to do so. The disadvantage of using a fringe projection system is the measurements are less accurate than alternative tactile methods - and typical methods to obtain an uncertainty evaluation within fringe projection require a tactile system as a comparator. Anterior to any measurement, fringe projection systems undergo a calibration, whereby the set of functional quantities (defined in this thesis as the system parameters) are found that define the measurement (the point-cloud) from the indication (a set of images). The accuracy of the estimated parameters will define the accuracy of any measurements made by the system. The calibration process does not evaluate any uncertainty of the estimated system parameters - the accuracy of the estimation of the parameters remains unknown, as is their exact effect on the measurement result. In this thesis, an investigation into the using the system parameters to evaluate the uncertainty of fringe projection measurements is made. Firstly, a method to localise the centre of ellipses in camera images with an uncertainty is given. This uncertainty is used to derive the uncertainty in the estimated system parameters. The uncertainty in the system parameters is tested by using the system parameters to measure known artefacts, a flatness artefact and two sphere-based artefacts, where the propagated uncertainty is tested against the measurement error. The accuracy of the system parameters are tested by comparing the measurement error of the measurements with measurements made on a commercial system, the GOM ATOS Core 300. In addition, an exhaustive study is undertaken on the calibration, including applying curvature, specificity and parameter stability tests on the non-linear regression used within calibration. The sphere-based measurements were found to not be robust enough against measurement noise in fringe projection to be able to provide information on errors caused by the system parameters. This thesis raises questions as to the appropriateness of using sphere-based measurements to represent the performance of a fringe projection system. The flatness measurements made using the estimated system parameters achieved an accuracy of approximately 30 "μm" across a 300 "mm"×140 "mm" flatness artefact, which is similar to measurements made by the commercial system. However, the estimated uncertainty was unable to explain all measurement discrepancy between the fringe projection measurements and the tactile measurements. The result specificity test indicated poor specificity of the mathematical model of fringe projection, namely the camera pinhole model with Brown-Conrady distortion. It is concluded that the level of accuracy of the mathematical model has become a limiting factor in the accuracy of fringe projection measurements, instead of the accuracy of the inputs to the calibration. Therefore, the uncertainty of the system parameters cannot be used to evaluate an uncertainty of a measurement made using a fringe projection system.

Inexact optimisation techniques such as heuristics and metaheuristics that quickly find near-optimal solutions are widely used to solve hard problems. While metaheuristics are well studied on specific problem domains such as travelling salesman, timetabling, vehicle routing etc., their extension to engineering domains is largely unexplored due to the requirement of domain expertise. In this thesis, we address a specific engineering domain: the design of sound-absorbing porous materials. Porous materials are foams, fibrous materials, woven and non-woven textiles, etc., that are widely used in automotive, aerospace and household applications to isolate and absorb noise to prevent equipment damage, protect hearing or ensure comfort. These materials constitute a significant amount of dead weight in aircraft and space applications, and choosing sub-optimal designs would lead to inefficiency and increased costs. By carefully choosing the material properties and shapes of these materials, favourable resonances can be created making it possible to improve absorption while also reducing weight. The optimisation problem structure is yet to be well-explored and not many comparison studies are available in this domain. This thesis aims to address the knowledge gap by analysing the performance of existing and novel heuristic and metaheuristic methods. Initially, the problem structure is explored by considering a one-dimensional layered sound package problem. Then, the challenging two-dimensional foam shape and topology optimisation is addressed. Topology optimisation involves optimally distributing a given volume of material in a design region such that a performance measure is maximised. Although extensive studies exist for the compliance minimisation problem domain, studies and comparisons on porous material problems are relatively rare. Firstly, a single objective absorption maximisation problem with a constraint on the weight is considered. Then a multi-objective problem of simultaneously maximising absorption and minimising weight is considered. The unique nature of the topology optimisation problem allows it to be solved using combinatorial or continuous, gradient or non-gradient methods. In this work, several optimisation methods are studied, including solid isotropic material with penalisation (SIMP), hill climbing, constructive heuristics, genetic algorithms, tabu search, co-variance matrix adaptation evolution strategy (CMA-ES), differential evolution, non-dominated sorting genetic algorithm (NSGA-II) and hybrid strategies. These approaches are tested on a benchmark of seven acoustics problem instances. The results are used to extract domain-specific insights. The findings highlight that the problem domain is rich with unique varieties of solutions, and by using domain-specific insights, one can design hybrid gradient and non-gradient methods that consistently outperform state-of-the-art ones.

Novel, porous and dense, ceramic (Fe3O4 & Ca2Fe2O5) and glass-ceramic (phosphate-based glass matrix with magnetic domains) magnetic microspheres have been manufactured for the first time via a rapid, single-stage, flame spheroidisation process. Morphological, structural, and compositional investigations provide evidence into the microsphere formation mechanisms, as a function of Fe3O4 precursor particle size, precursor-to-porogen mass ratio, and gas flow settings. Optimised conditions for the flame spheroidisation processing of these ceramic and glass-ceramic, porous and dense magnetic microspheres are identified. The magnetic properties of the materials provided for controlled induction heating to a constant level (40 - 45 °C), making these microspheres highly appropriate for localised magnetic hyperthermia applications. Complementary, cytocompatibility investigations confirmed the suitability of porous microspheres for biomedical applications. It is suggested that the flame-spheroidised materials developed opens up new opportunities for the rapid manufacture of innovative synergistic biomaterials, towards magnetic hyperthermia applications.

Ice accretion often posed serious operational and safety challenges in a wide range of industries, such as aircraft, wind turbines, power transmission cables, oil field exploration and production and marine transport. Great efforts had been expended to research and develop viable solutions for both active and passive ice prevention. Effective ice protection techniques, however, had yet to be developed due to their deficiencies, such as significant consumption of de-icing chemicals or energy for active methods and poor balance among durable icephobic performance, good environmental resistances, and good mechanical and chemical properties for passive methods. Thus, with the specific surface microstructure design and the deployment of interdisciplinary techniques, this PhD project aimed to research and develop highly efficient ice mitigating strategies. The outcomes of this project committed to widening the vision of the ice mitigating field and minimising the gap between lab research and practical application. Before the experimental work, a fundamental literature review was conducted about the intrinsic background of ice crystals and iced surfaces, the relationship between hydrophobicity and icephobicity, the current research status of ice mitigation, and the commonly used characterisation and evaluation of icephobicity. Inspired by the talent ideas from literature, the interdisciplinary techniques, powder metallurgy and surface acoustic wave (SAW), were chosen to develop new strategies that covered from dry passive textured surfaces and wet passive icephobic structures to active ice protection with high efficiency. Through a simple stamping method and chemical vapour deposition, the textured polydimethylsiloxane-184 (PDMS) and 1H,1H,2H,2H-perfluorooctyltriethoxysilane (POTS) -coated coatings with fine and regular texture patterns were developed successfully. The evaluations of surface wettability, water droplet icing, environmental scanning electron microscope (ESEM) icing, and ice adhesion strength demonstrated the good anti/de-icing performance of the prepared coatings. The roles of lowering surface energy and surface texturing for the improvement of hydrophobicity and icephobicity were also discussed and verified accordingly. The relationship between the ice adhesion strength and surface wetting factors gave poor correlations which indicated that the icephobic mechanisms of textured surfaces could not be correlated with single surface wetting parameters or by linear fitting. With the application of solid sintering, surface functionalisation, and liquid impregnation, two new wet strategies, liquid-impregnated porous metallic structure (LIPMS) and phase change materials (PCM)-impregnation porous metallic structures (PIPMSs), were proposed and fabricated with gradient porosity and hydrophobic/oleophilic guarding design. The evaluation of mechanical properties and repeated anti/de-icing performance showed effective icing delay, good humidity resistance, ultra-low ice adhesion strength (less than 1 kPa), good mechanical properties, durable icephobicity, and improved liquid-maintaining capacity for both LIPMS and PIPMS. Due to the solid properties of PCMs at subzero temperatures, PIPMS had better durable performance and impregnated-material maintaining capacity than LIPMSs. The surface localised acoustic waves and the acoustic thermal effect of the thin film zinc oxide (ZnO) surface acoustic wave devices demonstrated their excellent potential for active ice mitigations. With the specific surface design, SAW devices were proven to have good anti/de-icing performance with high energy efficiency in a series of active anti/de-icing experiments. The sensitive monitoring of the resonant frequency shifts and their good ice repellency proved SAW devices can be used to design an effective ice monitoring-removing autonomous system. The anti/de-icing work with rime icing on SAW devices revealed the dynamic interfacial behaviour of SAW that the hybrid effect of acoustic waves and the acoustic thermal effect led to the effective anti/de-icing performance.

Phosphate-based glasses (PBGs) are hugely promising materials for bone repair and regeneration as they can be formulated to be compositionally similar to the inorganic component of bone. Alterations to PBG formulations can be made to tailor their degradation rates and subsequent release of biotherapeutic ions to induce cellular responses, such as osteogenesis. In this work, novel invert-PBGs in the series xP2O5·(56-x)CaO·24MgO·20Na2O (mol%), where x is 40, 35, 32.5 and 30, were formulated to contain pyro (Q1) and orthophosphate (Q0) species. These PBGs were then processed into highly porous microspheres (PMS) via a flame spheroidisation process developed within the research group. Compositional and structural analysis using EDX and 31P-MAS NMR analysis revealed significant depolymerisation had occurred with reducing phosphate content, which increased further when PBGs were processed into PMS. A decrease from 50% to 0% of Q2 species and increase from 6% to 35% of Q0 species was observed for the PMS when the phosphate content decreased from 40 to 30 mol%. Ion release studies also revealed up to a 4-fold decrease in cations and an 8-fold decrease in phosphate anions released with decreasing phosphate content. In vitro bioactivity studies revealed that the orthophosphate rich PMS had favourable bioactivity responses after 28 days of immersion in SBF. Indirect and direct cell culture studies confirmed that the PMS were cytocompatible and supported cell growth and proliferation over 7 days of culture. The P30 PMS with ~65% pyro and ~35% ortho phosphate content revealed the most favourable properties and was proposed to be highly suitable for bone repair and regeneration, especially for orthobiologic applications owing to their highly porous morphology. Doxorubicin (DOX) was used as a model drug to assess its loading and release kinetics from porous phosphate-based glass microspheres to ascertain their suitability for localised drug delivery for the treatment of bone cancers. P40 PMS revealed a DOX loading efficiency of 55%, which was significantly greater than P30 PMS at 29.1%. Both P40 and P30 PMS released more DOX in phosphate buffered saline (PBS) at pH 5 as compared to release at pH 7.4. P40 PMS released 57% of DOX at pH 5 over a 48-hour period, whereas P30 PMS only released 15% of DOX. A pH-responsive DOX release in a more acidic environment suggests that the chemotherapeutic delivery and efficacy properties may lead to increased drug release within tumour tissues. Internal radiotherapy has been shown to be an effective treatment modality to destroy cancerous tissues and is usually achieved by the placement of radioactive sources at the tumour site. In this work, a novel processing method was established to combine yttrium oxide (Y2O3) with P40 phosphate glass particles to form uniform, solid microspheres containing very high yttrium levels via our flame spheroidisation process. The 30Y (~15 mol% Y2O3) and 50Y microspheres (~39 mol% Y2O3) had equivalent and superior yttrium content in comparison to clinically available microspheres used for internal radiotherapy (i.e., Therasphere®). The yttrium-containing microspheres formed were shown to be glass-ceramics, with crystalline phases present but with all elements homogenously distributed throughout the microspheres. Increasing yttrium addition resulted in increased durability of the microspheres, with 50Y microspheres revealing a 10-fold decrease in the release rate of some ions compared to P40 solid microspheres. Indirect and direct cell culture studies confirmed that the 30Y and 50Y microspheres were cytocompatible and supported cell growth and proliferation over 7 days of culture. No significant difference was observed in the metabolic and ALP activity for MG63s for both 30Y and 50Y microspheres from both indirect and direct cell culture studies. Yttrium was incorporated into the phosphate-based microspheres at a level that had not previously been achieved or observed from the literature studies and were shown to support bone cell attachment and growth. A high yttrium content could enable more radiation to be delivered per dose of microspheres, resulting in shorter neutron activation times which could prove beneficial for logistical issues associated with transportation of the biomaterials following nuclear activation. The radionuclide holmium-166 (166Ho) which is comparable to yttrium-90 (90Y) in that it emits β-radiation with a similar tissue penetration range and a significantly reduced half-life of 26.8 hours, was also investigated. The beneficial paramagnetic properties and density of 166Ho indicates that 166Ho-doped materials could be visualised through clinical imaging techniques, whilst simultaneously delivering a therapeutic dose of radiation. In this work, solid holmium-containing microspheres were similarly produced via the flame spheroidisation process using holmium oxide (Ho2O3) and P40 phosphate glass particles. The glass-ceramic microspheres produced had equivalent (30H: ~17mol% Ho2O3) and superior (50H: ~30mol% Ho2O3) holmium content in comparison to clinically used yttrium-doped microspheres (i.e. Therasphere®). Analogous to yttrium containing microspheres, elevated holmium content resulted in topographically unique features on the surface of some 50H microspheres. This increased holmium content resulted in significantly reduced ion release rates for all the ions and the holmium-microspheres did not show evidence of bioactivity. However, in vitro indirect and direct cell culture studies demonstrated their cytocompatibility. No significant difference was observed in the metabolic and ALP activity of MG63 cells for 30H and 50H microspheres in both the indirect and direct cell culture methods. This study appears to be the first to demonstrate microspheres containing high levels of holmium content that can also facilitate direct cell growth and proliferation of human osteoblast-like cells. The microspheres developed are therefore hugely promising biomaterials for both drug delivery and internal radiotherapy applications, as well as for promoting bone repair and regeneration at damaged sites. High holmium content could also result in higher specific activity per microsphere to increase radiotherapy delivery whilst also promoting higher visibility via imaging modalities.

The accumulation of ice on critical infrastructure, such as aviation, telecommunication, electricity, and transportation systems, can cause severe disruptions to daily life and significant economic losses. While applying icephobic coatings has been proposed as a potential solution, the challenge remains in creating coatings that can effectively withstand adverse environmental conditions, and also demonstrate a long service life and mechanical durability. The motivation of the work is to provide a comprehensive analysis of the fundamental aspects related to icephobicity research, and to explore novel, durable icephobic materials and surfaces, to produce well-rounded research outcomes. Liquid-infused surfaces (LIS) with self-replenishing properties have been adopted for anti-icing and de-icing applications. However, the mechanical durability of the LIS polymer is limited, rendering them unsuitable for long-term practical use in ice protection. To alleviate the difficulties associated with limited mechanical durability and rapid oil depletion among LIS structures, this study proposes a multi-phase LIS construction based on microporous metallic scaffolds. The proposed approach aims to tackle the problem of icephobicity deterioration in LIS materials. Specifically, the study employed Ni scaffolds and a PDMS-modified silicone oil (NP-SO) system for the LIS construction. The inclusion of Ni scaffolds facilitated stress redistribution during de-icing, thereby safeguarding the LIS material from external loads and impacts. As a result, the ice adhesion strength of the NP-SO samples down to 2.0 ± 0.7 kPa after undergoing 50 cycles of anti-icing/de-icing tests. Conversely, samples without Ni scaffolds demonstrated surface damage after 30 cycles of anti-icing/de-icing tests. The rapid depletion of infused oil due to large polymer deformation was also addressed. The multi-phase construction reduced oil depletion by over three times after 50 cycles of anti-icing/de-icing tests, providing a resolution for the oil depletion concerns in extended service environments. To achieve enhanced icephobic performance and durability, a novel design concept involving the integration of phase change liquids with Ni scaffolds and polydimethylsiloxane (PDMS) has been introduced in the multi-phase icephobic construction. Incorporating phase change liquids within the LIS construction allows for heat absorption and release during the icing process, which enhances icephobic performance. Differential scanning calorimetry analysis demonstrated that samples with phase change liquids have a phase change enthalpy of 3.2 J/g for groundnut oil and 3.8 J/g for coconut oil, respectively. As a result of the heat release during the phase change process of the incorporated phase change liquids, the surface temperature of the sample increased, resulting in a delayed ice formation of the supercooled water droplet. The observed freezing time of the droplets on the sample surface was 7-8 times longer than that observed on aluminum alloy. Solidification of the phase change liquids occurred at a low temperature, leading to a concurrent reduction in oil depletion during the de-icing process. This effect was confirmed from the weight maintenance ratio of samples, which remained unchanged even after undergoing 50 cycles of anti-icing/de-icing tests. The introduction of phase change liquids has been demonstrated to be a viable strategy for enhancing the anti-icing performance of icephobic surfaces. A high-quality icephobic surface should encompass multiple facets of icephobicity, such as anti-icing performance, de-icing ability, and mechanical stability. It is noteworthy that achieving all of these requirements simultaneously is a formidable challenge. To address this issue, a potential icephobic structure has been proposed, which incorporates Ni scaffolds with ice depressing liquid and a polydimethylsiloxane (PDMS) matrix. The ice depressing liquid utilized in this study included glycol and glycerol. The incorporation of ice depressing liquid into the surface layer led to an increase in the amount of hydroxyl groups present, resulting in a notable reduction in the freezing point of supercooled droplets at the liquid-solid interfaces. Furthermore, samples containing ice depressing liquid exhibited a considerable delay in the onset of icing, with respective times of 153.7 ± 2.5 s and 209.7 ± 4.5 s. The results of anti-icing tests provided evidence of the efficacy of ice depressing liquid in reducing the icing point temperature. Upon the sample preparation, the glycol and glycerol in the ice depressing liquid underwent volatilization, leading to the formation of internal porosities within the PDMS. Despite this, the surface morphology of the samples remained smooth and intact. The presence of these internal pores could increase the difference in elastic modulus between the polymer and metallic scaffolds, thus facilitating the initiation of micro-cracks at the ice-solid interfaces and aiding in ice detachment from the sample surface. Consequently, the ice adhesion strength was reduced to below 2 kPa, in contrast to 29.8 ± 2.9 kPa observed for Ni-PDMS samples. Moreover, the results indicated that the de-icing ability of the samples remained stable throughout cyclic icing/de-icing tests.

Additive Manufacturing (AM) sits at the forefront of technologies that offer unprecedented capabilities to lead a new industrial revolution. The higher degrees of complexity and flexibility in manufacturing intricate structures whilst maintaining accuracy and precision is one of the key selling points of AM technologies. Drop-on-demand metal jetting (DoD-MJ) is a promising additive manufacturing technique, gaining interest for its direct high-resolution printing of complex single and multi-material components. It also has key advantages over other metal AM techniques, such as avoiding powder handling and extensive post-processing. DoD-MJ processes are categorised based on the droplet generation actuation method. MagnetoHydroDynamic (MHD) actuators are the most advanced devices to date that have overcome the challenges of producing high-temperature droplets at high rates. MetalJet, the MHD-based system used in this thesis, has the capacity to produce molten micro-droplets (60-90 μm) at temperatures up to 2000 °C to form single and multi-material objects at 500 Hz. Applications for this technology include flexible circuits, advanced electronic components, intricate 3D mesoscale objects, and biotechnologies. In this method, 3D parts are built via spatially controlled deposition of individual molten droplets onto a substrate. Therefore, the success of the process entirely depends on the behaviour of these single droplets from deposition to solidification. Owing to its novelty, further utilization of this technology is impeded to date by a lack of understanding of various aspects of the process, including droplet generation and interface formation, residual stress development and microstructure evolution, which have been scarcely investigated to date. This thesis uses an integrated numerical, experimental, and analytical approach to provide insights into these research questions. The in-house MetalJet platform was used to investigate formation, spreading, and solidification of metallic micro-droplets at low Weber numbers. This was undertaken by ejection and deposition of single and multiple droplets of low (Sn) and high-temperature (Ag and Cu) metal onto various substrates using a range of jetting and substrate temperatures. High-speed photography provided insight into the mechanism of droplet generation. Analytical modelling based on the scaling method was performed to investigate the dynamics of droplets on deposition. Thermal Finite Element (FE) models were used alongside to explain the experimentally-observed morphology of droplets and investigate the droplet-to-substrate and droplet-droplet adhesion. Moreover, thermo-mechanical FE modelling was used to study the residual stress evolution during the solidification and cooling processes, and investigate its role in the observed droplet warping and delamination. Furthermore, various characterisation techniques, including FIB-SEM, EDX, EBSD, and TEM, were used to evaluate the microstructure of printed droplets. For the first time, the generation mechanism of uniform microdroplets at high temperatures and the role of jetting voltages are explained in this thesis. Moreover, this study reports that increasing the substrate temperature intensified the diffusion between the droplet and substrate, which resulted in improved adhesion. It is also shown that ripples forming on a droplet's periphery during solidification disappeared at elevated substrate temperatures, resulting in enhanced inter-droplet bonding. Furthermore, the significant role of the substrate wettability and thermal properties, which control the droplet's dynamics and solidification behaviour, respectively, is elucidated. This highlights the importance of the substrate material selection for this technology. It is demonstrated that the steep temperature gradients during deposition result in residual stresses, and if they exceed the elastic limits of the material, stress-induced delaminating can occur. Finally, results presented in this thesis underpin the optimal process conditions under which the 3D structures produced with this technology can exhibit reliable integrity and consistency. Overall, the works represent a step forward towards the direct metal printing of high-resolution functional multi-material components.

Vehicular networks have been identified as a key enabler for future smart traffic applications aiming to improve on-road safety, increase road traffic efficiency, or provide advanced infotainment services to improve on-board comfort. However, the requirements of smart traffic applications also place demands on vehicular networks' quality in terms of high data rates, low latency, and reliability, while simultaneously meeting the challenges of sustainability, green network development goals and energy efficiency. The advances in vehicular communication technologies combined with the peculiar characteristics of vehicular networks have brought challenges to traditional networking solutions designed around fixed parameters using complex mathematical optimisation. These challenges necessitate greater intelligence to be embedded in vehicular networks to realise adaptive network optimisation. As such, one promising solution is the use of Machine Learning (ML) algorithms to extract hidden patterns from collected data thus formulating adaptive network optimisation solutions with strong generalisation capabilities. In this thesis, an overview of the underlying technologies, applications, and characteristics of vehicular networks is presented, followed by the motivation of using ML and a general introduction of ML background. Additionally, a literature review of ML applications in vehicular networks is also presented drawing on the state-of-the-art of ML technology adoption. Three key challenging research topics have been identified centred around network optimisation and ML deployment aspects. The first research question and contribution focus on mobile Handover (HO) optimisation as vehicles pass between base stations; a Deep Reinforcement Learning (DRL) handover algorithm is proposed and evaluated against the currently deployed method. Simulation results suggest that the proposed algorithm can guarantee optimal HO decision in a realistic simulation setup. The second contribution explores distributed radio resource management optimisation. Two versions of a Federated Learning (FL) enhanced DRL algorithm are proposed and evaluated against other state-of-the-art ML solutions. Simulation results suggest that the proposed solution outperformed other benchmarks in overall resource utilisation efficiency, especially in generalisation scenarios. The third contribution looks at energy efficiency optimisation on the network side considering a backdrop of sustainability and green networking. A cell switching algorithm was developed based on a Graph Neural Network (GNN) model and the proposed energy efficiency scheme is able to achieve almost 95% of the metric normalised energy efficiency compared against the "ideal" optimal energy efficiency benchmark and is capable of being applied in many more general network configurations compared with the state-of-the-art ML benchmark.

Road transport contributes approximately 20% to the United Kingdom's greenhouse gas emissions, accelerating the effects of global warming. Since the United Kingdom, like many other countries, has pledged to reach net zero carbon emissions over the next two decades, reducing emissions from road vehicles has become a priority. A further adverse effect of road vehicle emissions is their link to serious health issues such as respiratory and cardiovascular diseases. To achieve the required substantial reduction in emissions, a multi-faceted approach will be required. In this project, one important aspect, the aerodynamics of ground vehicle platoons, is explored with the aim of expanding the understanding of road vehicle aerodynamics and exploring innovative solutions to improve road vehicle efficiency. Vehicle platooning is a form of cooperative travelling in which vehicles drive closely together, with the intention to reduce overall air resistance, fuel consumption and vehicle emissions. Platooning, i.e., the cooperative movement of a group of individuals, is a concept that is not unique to road vehicles, but can be commonly observed in nature (e.g., a school of fish) or in sport (e.g., cyclists riding their bikes in a train). Here the trailing individuals take advantage of the sheltering provided by the leading individuals of the group. As a continuation of this observation, it would be natural to assume that road platooning is always beneficial, and that the trailing vehicles of a platoon reliably experience a reduction in drag. However, there are several examples in the literature that report a rear vehicle in a platoon receiving a drag increase. With the wide range of vehicle geometries on the roads, it is vital that additional research is targeted at understand the fundamental aerodynamic principles that lead to such adverse platooning results and understand the role that geometry plays in influencing the effectiveness of a platoon. In the first stage of this project, the geometry dependence of platooning was explored by systematically altering the shape of a simplified ground vehicle to change its platooning behaviour from the 'classical' platooning behaviour, where the rear vehicle experiences a high drag reduction, to 'inverted' platooning behaviour, where the rear vehicle suffers an increase in drag. To this end, a large parameter study was completed using unsteady Reynolds-Averaged Navier-Stokes (URANS) simulations. A key outcome of this study was that the combination of a more streamlined rear vehicle, coupled with strong wake-impingement caused by the lead vehicle results in the most adverse platooning outcome. The second stage of the project focused on establishing the potential of using passive flow control to alleviate the adverse platooning effects that were observed in a platoon composed of two Ahmed bodies with 25◦ rear slant angles. First, the potential of plasma actuators as flow control devices was explored by experimentally characterising the performance of a serrated dielectric barrier discharge (DBD) plasma actuator. This was followed by another set of URANS simulations which considered the application of flow control in the context of a platoon of two 25◦ Ahmed vehicles. This covered both plasma-actuator like induced jets as well as flaps as flow control devices. The flow control devices were located at the top of the rear slant of the front vehicle and were designed to induce flow separation to increase the size of the front vehicle's wake. Using this technique a drag reduction for the rear vehicle of up to 25% compared to the configuration without flow control was achieved. In the final stage, the effectiveness of flow control was tested experimentally in the University of Glasgow's Handley-Page wind tunnel. First the dependency of the drag coefficient of a platoon composed of two 25◦ Ahmed vehicles on inter-vehicle spacing and Reynolds number was investigated, showing that a significant dependency on both parameters exists. Then, flow control was introduced in the form of a flap, with the previous sets of experiments being repeated for three flap angles and two flap lengths. While the flap was not quite as effective as predicted by the URANS simulations, the flap still induced a significant reduction in drag (ca. 9%) when compared to the rear vehicle of the baseline case that was subject to inverted platooning conditions.

This thesis summarises the work performed in the use of flight simulation to assess the impact of wake encounters and turbulence on the safety of helicopter operations. An initial literature search is presented, revealing a lack of previous research regarding the impact of wind turbine wakes on helicopter operations and a lack of accurate and easy to implement modelling tools for the assessment of the impact of turbulence on rotorcraft handling. The work is divided in two main parts: In the first part, following a literature review, results of flight simulation tests representing the accidental crossing of a wind turbine wake by a helicopter are presented. Encounters are of a 'mild' severity, flight safety was not compromised and resulted mainly in excursions in the yaw axis and deviations in roll. The pilot workload required in the recovery of from the initial encounter was strongly related to the aircraft's handling qualities, especially yaw to pitch and yaw to roll cross couplings. The second part of the thesis details the development and implementation of a new turbulence modelling method for flight simulation. The model is based on an adaptation of a synthetic eddy method (SEM). It generates a random turbulence field surrounding the aircraft by filling a control volume with turbulence generating eddies which are displaced by ambient flow; this is the first time this type of modelling has been implemented in real-time piloted flight simulation. The induced turbulence automatically cross correlates disturbances across all aircraft elements and can be adjusted by changing strength, shape and size of eddies and combining series of eddies with different properties. Average frequency of the induced turbulence is proportional to the cube root of the number of eddies; increasing the frequency results in increased computational costs. Nevertheless, the SEM turbulence generator can produce real time disturbances within the 0.1 - 1Hz frequency range which has the most impact on pilot workload. Offline and piloted simulation was used to evaluate the model and compare against a precomputed wake using a Mann turbulence model. Results show that SEM induced turbulence can impact handling in similar manners as the precomputed wake, while offering random turbulence in real time.

The development of wearable sensors has become a major area of interest due to their wide range of promising applications, including health monitoring, human motion detection, human-machine interfaces, electronic skin and soft robotics. Particularly, pressure sensors have attracted considerable attention in wearable applications. However, traditional pressure sensing systems are using rigid sensors to detect the human motions. Lightweight and flexible pressure sensors are required to improve the comfortability of devices. Furthermore, in comparison with conventional sensing techniques without smart algorithm, machine learning-assisted wearable systems are capable of intelligently analysing data for classification or prediction purposes, making the system 'smarter' for more demanding tasks. Therefore, combining flexible pressure sensors and machine learning is a promising method to deal with human motion recognition. This thesis focuses on fabricating flexible pressure sensors and developing wearable applications to recognize human gestures. Firstly, a comprehensive literature review was conducted, including current state-of-the-art on pressure sensing techniques and machine learning algorithms. Secondly, a piezoelectric smart wristband was developed to distinguish finger typing movements. Three machine learning algorithms, K Nearest Neighbour (KNN), Decision Tree (DT) and Support Vector Machine (SVM), were used to classify the movement of different fingers. The SVM algorithm outperformed other classifiers with an overall accuracy of 98.67% and 100% when processing raw data and extracted features. Thirdly, a piezoresistive wristband was fabricated based on a flake-sphere composite configuration in which reduced graphene oxide fragments are doped with polystyrene spheres to achieve both high sensitivity and flexibility. The flexible wristband measured the pressure distribution around the wrist for accurate and comfortable hand gesture classification. The intelligent wristband was able to classify 12 hand gestures with 96.33% accuracy for five participants using a machine learning algorithm. Moreover, for demonstrating the practical applications of the proposed method, a realtime system was developed to control a robotic hand according to the classification results. Finally, this thesis also demonstrates an intelligent piezoresistive sensor to recognize different throat movements during pronunciation. The piezoresistive sensor was fabricated using two PolyDimethylsiloxane (PDMS) layers that were coated with silver nanowires and reduced graphene oxide films, where the microstructures were fabricated by the polystyrene spheres between the layers. The highly sensitive sensor was able to distinguish throat vibrations from five different spoken words with an accuracy of 96% using the artificial neural network algorithm.

The use of aluminium alloys in the construction industry is rising due to their profound features including high strength-to-weight ratio, good corrosion resistance, ease of processing, low maintenance, high recyclability and aesthetic appearance. However, the low modulus of elasticity, welding difficulty and low melting point of aluminium have adverse effects on the performance of structural members made of aluminium alloys. In case of aluminium alloy tubular members, the structural performance can be improved with the addition of concrete infill, despite the different expansion of aluminium alloy and concrete. In concrete-filled aluminium tubular (CFAT) structural members, the aluminium tube increases the compressive performance of the concrete core due to the confinement effect, while the concrete core delays inward local buckling of the aluminium tube. Moreover, the self-weight of these structural members can be decreased further by replacing the inner concrete core with a hollow tube. Concrete-filled double skin aluminium alloy tubular (CFDSAT) member is made with two aluminium tubes and concrete infill between them. CFDSAT members retain all advantages of CFAT ones and additionally they have less self-weight. Moreover, these members offer better local and global stability because of the interaction of three components. Therefore, compared to steel-concrete composite members, CFDSAT members can be more efficient for structures situated in offshore areas and seismic-prone regions. However, the research on the structural performance of aluminium alloy-concrete composite members is minimal. Therefore, this study investigates the behaviour of CFAT and CFDSAT structural members subjected to axial compression and bending. In the experimental programme of CFAT structural members, a total of 18 columns, including 9 CFAT and 9 bare aluminium tubular (BAT) specimens and 20 beams, including 10 CFAT and 10 BAT specimens are tested. The BAT structural members are tested for reference purposes and to assess the applicability of Eurocode 9 design standard. The experimental investigation of CFDSAT structural members includes 8 columns and 10 beams. The column tests are conducted using a pin-ended set-up for allowing the specimens to rotate around the buckling axis. The flexural members are tested in a four-point bending arrangement to spread out the maximum stress over the area between the two loading points. The material properties of aluminium alloy are determined by tensile coupon tests. The structural responses obtained from the experiments are presented in terms of ultimate capacity, failure modes and load/moment versus mid-length defections curves. Finite element (FE) models of the structural members are developed by taking into account the geometric and material nonlinearities and validated against the experimental results. The validated models are adopted to conduct parametric studies to examine the effects of different design parameters on the behaviour of the structural members. In the absence of design rules for aluminium alloy-concrete composite structural members, design methodologies are proposed to predict the ultimate capacity of these composite members based on the Eurocode 4 framework. To obtain information about the sustainability of aluminium alloy-concrete composite structural members, the life-cycle performance of CFAT and CFDSAT columns is investigated. Life-cycle assessment and life-cycle cost analysis methods are applied to evaluate the long-term environmental and economic aspects of these members, respectively. The cradle-to-grave system boundary is considered for these analyses to cover all the aspects of life-cycle. Finally, a comparison of the self-weight of these members is also presented. The experimental results show that compared to the bare specimens, the counterpart composite specimens have remarkably improved ultimate capacity, stiffness and ductility due to the concrete infill and the improvement is more pronounced for the specimens with thinner sections. The FE parametric investigation reveals that the larger cross-sectional dimensions of the outer tube of composite members substantially improved the load-bearing capacity, while the cross-sectional dimensions of the inner tube have a negligible effect. Moreover, FE results show that higher grade concrete remarkably increased the strength of composite columns, while it has a less significant influence on the improvement of the flexural strength of composite beams. In the absence of design standards for aluminium alloy-concrete composite structural members, design methodologies are provided based on Eurocode 4 (2004) to determine the strength of these members. The self-weight comparison and sustainability study results indicate that CFDSAT structural member is lighter, while the CFAT member is less expensive and provides the lowest environmental impact than the other alternatives considered in this study.

This research aims to develop agro-waste based unfired earth blocks that are suitable for tropical climate. The samples were produced using eggshell (10, 20, 30, 40, and 50wt.%), sawdust (2.5, 5, 7.5, and 10wt.%), coconut husk (2.5, 5, 7.5, and 10wt.%), and walnut shell (5, 10, 15, and 20wt.%) separately and in combination. In this study, several physico-mechanical, durability, and thermal properties tests were conducted. Moreover, numerical computer simulation analyses on the thermal performance of the developed materials were performed. Additionally, the effect of three different particle sizes of sawdust (212 μm ‹ x ‹ 300 μm, 425 μm ‹ x ‹ 600 μm, and 1.18 mm ‹ x ‹ 2.00 mm) was investigated. The results of physico-mechanical properties tests showed that the density of the samples gradually decreased with increasing agro-wastes percentages in the mixture. Besides, eggshell and walnut shell addition reduced the linear shrinkage and capillary water absorption, whereas finer sawdust particles and coconut husk incorporation increased both. The individual inclusion of eggshell, sawdust, and coconut husk improved the strength of clay blocks; however, walnut shell had a negative impact on the strength. On the other hand, combining eggshell with other agro-wastes did not contribute to improving the strength. The highest compressive strength (CS) and flexural strength (FS) of air-dried samples were obtained when 40% eggshell (CS: 5.68 MPa, FS: 2.24 MPa), 2.5% coconut husk (CS: 4.78 MPa, FS: 2.14 MPa), and 2.5% sawdust (CS: 4.74 MPa, FS: 2.00 MPa) were individually used. The compressive strength of eggshell, coconut husk, and sawdust samples improved by around 40%, 18%, and 17%, respectively, while the flexural strength enhanced about 47%, 41%, and 32% over the reference sample. The investigation of durability tests revealed that when agro-wastes were added, erosion resistance improved compared to the reference sample. According to the thermal properties tests, thermal conductivity and volumetric heat capacity decreased with the addition of agro-wastes. The lowest conductivity was obtained for the 10% coconut husk sample (0.172 W/mK), followed by the 10% sawdust (0.185 W/mK) and the 50% eggshell (0.299 W/mK) samples. Furthermore, in the wall scale test, the sample walls made of coconut husk, sawdust, and eggshell outperformed the reference sample wall in terms of thermal resistance, with improvements of around 48%, 35%, and 16%, respectively. This result was also supported by the numerical thermal simulation analyses, where the findings showed that coconut husk wall construction had a relatively lower (14.90 MWh) yearly energy consumption than sawdust (15.10 MWh), eggshell (17.70 MWh), and reference (18.90 MWh) walls. The overall findings of this study suggest that the incorporation of agro-wastes into clay blocks improved their properties, making them suitable for use as building construction material.

Perceiving the 6D pose of object is a longstanding question. It plays a crucial role in some areas of robotics, such as object manipulation, grasping, unmanned aerial vehicles and autonomous vehicles. Although researchers have proposed various algorithms to address this problem in the history, like template matching, point pair feature, etc. In this thesis, we utilise deep learning techniques to overcome the limitations of some existing pose estimation algorithms. Specifically, we investigate two different tasks in perceiving the orientation and translation of an object in 3D space, pose estimation from single images, and pose tracking from video sequences. For the pose estimation task from single image, we introduce a novel channel-spatial attention network, which can learn the representative features from RGB-D images. Although there are some supervised Convolutional Neutral Network (CNN) frameworks used for estimating the object pose, they simply fuse the image features and geometry features together, which result in weak representations of fusion data as these multimodal data lays in various feature spaces. To address this, our channel-spatial attention network proposes a specific CNN that learns the most important embedding from each data format, and convert them to the identical dimensional feature space. To extend the content of the pose estimation task from single image, in this thesis, the pose tracking task is also investigated. We propose a novel tracking framework that achieves stable and real-time tracking in the process. This framework is based on the correspondence of two consecutive frames, where the temporal-spatial information is utilize. In summary, the goal of this thesis is to extend the research direction of pose estimation tasks in 3D space, especially by introducing some advanced deep learning techniques to this area. In this thesis, it shows that our deep learning based methods have the advantages of dealing with occlusions, and cutter background images over some existing object pose estimation methods.

Masonry walls are vulnerable to extreme loading and fibre-reinforced polymer (FRP) composites are a good solution for the reinforcement of un-reinforced masonry (URM) walls exposed to in-plane or out-of-plane loads. Collapse of walls in masonry structures is one of the main reasons for material damage and people's injury due to earthquakes. Thus, the investigation of effective and inexpensive methods for strengthening masonry walls is crucial. FRP composites have many benefits, such as a high strength-to-weight ratio, and ease of application in terms of workforce and time. The behaviour of masonry walls can be improved by retrofitting with fibre reinforced polymeric (FRP). Reinforcement of masonry walls with FRP, can increase strength and ductility and thus reduce the risk of failure under extreme loading. However, the reinforcement of cracked masonry walls and the interface and bond behaviour between brick and mortar, especially in the case of retrofitting with FRP, has not been carefully studied. The main aim of this research is to investigate the bond behaviour of different carbon fibre-reinforced polymer (CFRP) configurations on masonry structures when subjected to compressive and shear loads. Also, the experimental shear tests on cracked triplet masonry strengthened with different configurations of CFRP were carried out which had not been studied before. The experimental work was carried out with seven different CFRP configurations and layers on triplet masonry and each test was repeated four times. In addition, finite element modelling by the commercial software Ansys and verification against results from experimental tests were carried out. The present study has shown the strength of masonry triplets with CFRP under compression loads was increased up to 3 times and under shear loads, all types of reinforced triplets were at least 4 to 6 times stronger than triplets without CFRP. Comparing one and two layers of CFRP wrap or CFRP strips on triplet masonry showed that there is no significant bond improvement between one layer and two layers. Depending on CFRP configurations, two layers make triplets 5% to 15% stronger, which is not sufficient. The cracked triplet samples repaired with CFRP and tested under shear loading showed significant improvement in bond strength and the load capacity of triplet masonry increased by 3 to 5 times in comparison with samples without initial damage. Finally, based on test results to predict the shear strength of masonry triplets with and without CFRP new formulas for bond behaviour were developed. In addition, for an accurate finite element modeling (FE), these formulas were added as the new user-defined functions (UDF), in the commercial software Ansys. The outcome of this thesis improves the current knowledge on the use of CFRP to reinforce masonry walls with brick and mortar which will contribute to the understanding of the effect of CFRP on masonry structures. Conclusions are provided and future research needs in the area of masonry strengthening with CFRP are outlined.

3D printing of edible materials has seen a surge of interest within the last decade. However, there exists a shortage of printable materials and the understanding of parameters that must be controlled to facilitate successful print jobs that are appealing to consumers and reproducible to prevent waste and keep costs down. 3D printed objects are created one layer at a time, with subsequent layers deposited on top of the previous one. This requires the shape to set quickly and be able to support the weight of the new layers above it. Most research on the 3D printing of edible materials has focused on extruding pre-gelled samples which maintain their shape due to yield stress, with no sol-gel transition taking place. This work aims to address the gaps in the literature present surrounding hot-extrusion 3D printing. Hydrocolloid gels that gel rapidly, with a high gel strength and recovery that allows it to maintain its shape after printing are ideal for this method. This work examined agar and kappa-carrageenan (кC) as candidate hydrocolloids. Initially these gels had their thermal characteristics analysed through rheology and micro differential scanning calorimetry (DSC), as the gel temperatures were a key parameter for hot-extrusion 3D printing. The storage modulus (G') was also used to determine printability of the gels. Above their gel temperature, gels should have a low G' which should rapidly rise below the gel temperature. 3% w/v кC with 2% w/v vitamin B1 had a G' of around 100,000 Pascals and was found to be printable. 2% w/v agar was not printable with the printing setup being used. Hereafter, agar was no longer used. Texture profile analysis showed that 3D printed gels were weaker than cast gels, undergoing delamination at the weak semi-fused sites between printed layers. Release tests were carried out and the 3DP were found to release more vitamin B1 compared to cast gels (78% vs 66% after 6 hours at 37 ˚C) owing to the layering allowing water to penetrate more quickly into the gel network. Sunflower oil (SFO) was then added into the systems by creating oil-in-water emulsions and then dispersing the кC within them to produce emulsion gels. This gave a system with controllable energy content and the ability to encapsulate lipophilic molecules in a straightforward manner. Two emulsifiers were used, tween 20 (T20) and whey protein isolate (WPI). T20-stabilised emulsions underwent flocculation upon addition of the кC, but the WPI-stabilised emulsion gels did not. DSC showed that if the concentration of кC was fixed at 3% w/w in the aqueous phase, then addition of up to 40% w/w sunflower oil did not affect the gelling temperatures (36-37 ˚C). This was reflected in the printability not being affected by varying the oil phase from 0-40% w/w. Post-printing rheology and texture profile analysis found that increasing the oil fraction did not affect the tested mechanical characteristics of the 3D printed gels as they still failed along the semi-fused sites between layers. Finally, to build on this work the emulsion gels were tested for stability and it was found that the oil droplet size remained constant over 8 weeks. Cinnamaldehyde, a lipophilic molecule was added to the oil phase and release tests were carried out. Unlike the vitamin B1 there was no significant difference in the final amount of cinnamaldehyde released over six hours between cast and 3D printed кC-emulsion gels. Lastly, erioglaucine disodium salt (EDS) was added and co-release was carried out. The EDS was found not to interfere with the release of the cinnamaldehyde and it released faster from the 3D printed gels compared to the cast gels. Overall кC-gels were found to be suitable for 3D printing under a narrow range of parameters with and without sunflower oil. The 3D printed gels perform differently to cast gels in terms of mechanical properties owing to the presence of semi-fused sites running throughout them. They can be utilised as customisable release vehicles that release hydrophiles at a faster rate and lipophiles at a similar rate compared to the equivalent cast gels.

The electrocardiogram (ECG) is a recording of the electrical activity of the heart and is widely used by clinicians as a diagnostic tool and during life-saving resuscitation efforts. The traditional 12-lead silver chloride ECG remains the gold standard in healthcare, despite promising advances in sensor technology and signal processing that have the potential to improve quality and ease of recording. The primary aim of this research is to improve the data acquisition and signal processing of the ECG. To achieve this, methods were developed to improve the utility and range of applications for non-contact electric potential sensors (EPS). These sensors allow for rapid, high resolution ECG recording and can be incorporated into passive monitoring systems and wearable devices. Dry flexible textile electrodes and hardware/software filtering systems were developed for use with EPS. These were characterised using simulated and pre-recorded ECGs, cardiac phantoms, and real world data through live human recording. Firstly, dry flexible textile electrodes were developed and characterised for EPS ECG sensing. These electrodes were shown to have the capacity to for the rapid acquisition of high quality ECG suitable for the calculation of heart rate, with an accuracy within ±1 beat per minute, and a 99% beat detection confidence compared to reference electrodes. Secondly, a novel adaptive filtering technique was developed and validated using a novel neonate phantom developed for EPS testing. Using the noise profile from the injection of a known signal, an algorithm for an adaptive ECG filter was generated. This filter was highly effective in removing the predominant mains power line noise that EPS sensors are particularly susceptible to, with average of 22 dB reduction. Additionally, this filtering method introduced no significant alteration of the ECG morphology, with a mean square error equal to that of traditional filtering techniques. This work advances the development of applications for EPS technology by demonstrating and characterising textile electrodes and novel filtering methods. Textile electrodes allow for the integration of these devices into convenient and comfortable devices, as explored in the proof of concept for a smart ECG sensing mattress for use in the neonatal intensive care unit. The novel filtering techniques displayed here are tailored to the characteristics of EPS sensors, and provide ECG signal quality that can compete with the high standards required in clinical practice.

Cracking of concrete and degradation are the major problems in today's concrete structures in the UK. Repair and maintenance of existing RC structures costs the UK government £50 billion per year. Most of these repairs (conventional repair) can last just 10 to 15 years. In addition to these technical aspects, the repair costs are usually very high. To achieve the sustainability targets, of the United Nations' Intergovernmental Panel on Climate Change (IPCC) to reduce global greenhouse gas (GHG) emission 40% - 70% by 2050, it is essential to reduce the environmental impact of reinforced concrete (RC) structures. Recyclable thermoplastic polymers might be a solution by filling the voids and/or reducing the conduction contact area for corrosive components of the environment. Since the root cause of most of the structural failure is attributed to concrete cracking, there is a compelling economic incentive to develop a concrete that can treat and repair the damage all by itself. Bacteria based self-healing concrete development, is relying on calcium carbonate (CaCO3) precipitation. The CaCO3 formed by bacteria in the cracks is a brittle material and mainly useful for healing of static cracks, the use of polymer will protect the bacteria in/ around the cracks and the combination can show a larger degree of elasticity that might allow to keep even a dynamic crack sealed. The goal of this research was developing self-healing concrete composites by immobilising bio-agents into the polymers. The use of bacteria-based solutions to complement the polymer modified concrete might be an opportunity by potentially enabling the re-use or increase the quantity of polymers incorporated and simultaneously taking advantage of polymers' elastic/ductile properties to heal concrete cracks in a complementary way to the calcium carbonate formed due to bacteria metabolism. This is the hypothesis of this work. The research started by analysing the influence of polymers on the mechanical and durability properties of concrete. A total of seven types of thermoplastic polymers were selected for this study, that they have not been considered before in the durability of infrastructures. The polymers are: ε-caprolactone (PCL), Polymorph, low density polyethylene copolymer with Ethylene-vinyl acetate (LDPE/EVA) granular and powder, general polyethylene (GPE), plastomer with ethylene-vinyl acetate EVA (N-228) and thermoplastic elastomers (TPE) (SEEPS); they were used as fine aggregate replacement (3 and 5%) and as a bio-agent carrier in concrete. Initially the experimental work was conducted to investigate the influence of the polymers and their replacement percentages on the mechanical and durability properties of concrete. As a result, the contribution of polymers leads to a 2-15% concrete strength increase. The durability properties such as open porosity, capillary water absorption improved for all the polymers except for TPE inclusion. For most of the polymers studied, their effect on preventing water migration via capillarity seems to occur 2-3 days after exposing concrete to water. The substitution of natural fine aggregates with some polymers leads to a reduction of chloride ion migration into the concrete samples, indicating that some of them stop free chlorides inside concrete. 5% LDPE/EVA leads to the higher restricted movement of free chloride migration as the coefficient decreased by 64% in comparison to plain concrete. This research then compares the performance of the three types of polymer modified concrete (TPE, LDPE/EVA, and PCL) if bacteria-based solutions are used. S. oneidensis MR-1 is a facultative anaerobic iron respiratory bacteria (IRB) that was selected to be investigated in this study. Experiments on the fresh, mechanical and durability characteristics of polymer and bio-polymer modified concrete, as well as the impact of nutritional media (Tryptic Soy Broth (TSB)) quantity (0.5L and 1L) were conducted. As a result, the bio-polymer (bacteria immobilised polymer) with 0.5L nutritional medium performed better when compared to 1L. Polymers and the bio-agent contributed to an increase in overall compressive strength, Bio-TPE and bio-LDPE/EVA exhibited a compressive strength of 75 MPa at 90 days. The contribution of bio-agent reduced concrete porosity from 12% to 2-4%. The water absorption via capillary was reduced by 85% in bio-polymer modified samples when compared to plain concrete. The chloride ion transfer was significantly reduced with the addition of polymers and the bio-agent, the improvement was 55 to 67% for 0.5 and 1L TSB at all ages. Surface electrical resistivity for bio-polymer modified concrete improved by 66%. According to the experimental outcomes, S. oneidensis could effectively induce calcium precipitation in the pH above 12. Through 16S rRNA gene sequencing it was confirmed that S. oneidensis can remain dormant up to 3-years in concrete. The maximum crack healed was 1.84 mm at 7 days, and completely fill the surface pores of 7.5 mm at 28 days after curing in water and TSB. Using SEM-EDS and FT-IR analyses, the healing material produced on or inside the cracks was investigated, and it was determined to be CaCO3.

In many engineering sectors with the potential to produce major accident hazards in the event of a failure, accepted good practice for implementing functional safety of Safety-Critical Systems (SCS) is considered conformity to the standard IEC 61508. Major accident hazards may potentially occur in the United Kingdom Commercial Marine Sector (UKCMS) in the event of a collision or capsizing of a large vessel, however IEC 61508 is not implemented sector wide. In this thesis, the International Maritime Organizations (IMO) Formal Safety Assessment is utilised as the framework for a model that determines a series of objectives required to overcome barriers to the implementation of IEC 61508 in the UKCMS. The typical cost-benefit analysis used during a Formal Safety Assessment is enhanced with a Balanced Scorecard (BSC) that assesses the barriers to the standards implementation from perspectives other than financial. The objectives and measures of the BSC are determined using a survey answered by engineers with a range of backgrounds and experience related both to functional safety of SCS in general, and the UKCMS. The rank in terms of importance of each objective, and of each perspective is determined using a fuzzy variant of the Analytical Hierarchy Process (AHP), then the top 25 ranked objectives are chosen for further investigation. The current success of the objectives are determined by investigating their progress relative to their measures. Afterwards, further actions for progressing measures below a certain score threshold are determined. Recommendations for achieving these actions are derived from a survey with experts, which are then used to form a strategy for UKCMS companies to migrate from current practice for implementing functional safety of SCS, to practice that aligns with IEC 61508. Objectives used in the BSC align with the typical perspectives of a BSC, however this project's first survey indicated the necessity to address regulatory and cyber-security related barriers as well. The validity of the results is measured against their ability to represent the full breadth of the UKCMS, which is dependent on expert participation. The migration strategy provides actions suitable for the current situation regarding conformity to IEC 61508, however the methods used in this project require iteration for continuous use. The methods used in this project and the current iteration of the migration strategy provides a solution for implementing IEC 61508 without a legislative incentive, and may also be useful should the standard be mandated in the future by IMO or a classification society. This thesis concludes with a proposal for the migration strategies implementation and iteration.

This thesis represents a summary of the activities I performed throughout my PhD, mainly on quantum optics in the infrared. The field of quantum optics was revolutionised by the experimental demonstration of correlated photon-pair generation via the non linear interaction known as spontaneous parametric down-conversion (SPDC). Quantum-enhanced optical systems utilising the unique properties of these quantum sources have blossomed in recent years particularly in the fields of quantum communications and quantum metrology. Currently, most measurements taking advantage of these non-classical sources have narrowed their focus on the near-infrared to telecommunications spectral region (0.7 − 1.6 µm) with good reason. Going beyond this spectral range, at longer wavelengths, presents significant technological problems in both generation, manipulation and detection of mid-infrared radiation. Fundamental advantages such as reduced scattering and propagation losses in well-established fields presents a compelling case to investigate non-classical states within the mid-infrared window. This thesis details the challenges associated with delving into the 2 µm regime and demonstrates the realisation of a compact and robust quantum source of entangled photon-pairs using custom designed non-linear crystals and superconducting nanowire single-photon detectors (SNSPD). The demonstration of two-photon interference and polarisation entanglement at 2.1 µm provides a solution which could prove valuable in the implementation of future freespace daylight quantum communications and high sensitivity metrology. This thesis follows up on this work by addressing a promising application of 2 µm sources in the field of integrated photonics. The testing of the performance of a type of novel hollow-core nested antiresonant nodeless fibre (HC-NANF) with mid-infrared radiation indicates high polarisation purity and low attenuation properties. This technology may provide the radical solution to ensure the future capacities of modern communication networks are met. The proposal to employ the quantum properties of entangled photon states generated via SPDC to enhance detection in wavelengths far beyond the mid-infrared is a challenging concept. This thesis moves in the direction of the application of non-classical states for the metrological detection of terahertz (THz) radiation of which has garnered significant interest in countless areas of scientific endeavour in recent times. It tackles the enhanced phase estimation granted by using NOON states in metrology applications and takes the first step in utilising alternative detection strategies for measuring THz radiation. Traditional techniques for measuring or reconstructing THz electric fields rely on balanced detection using photodiodes with inherent shot-noise limitations on the system. Terahertz time-domain spectroscopy (THz-TDS) using electro-optic sampling (EOS) and ultrashort pulsed probes is typically employed to measure directly the electric field of THz radiation. This thesis reports on the first step in the direction of using non-classical states for THz detection demonstrating that electric fields can be measured with single-photon detectors using a squeezed vacuum as the optical probe. The approach achieves THz electro-optical sampling using phase-locked single-photon detectors at the shot-noise limit and thus paves the way toward quantum-enhanced THz sensing.

Future large space structures are likely to be constructed at much greater length-scales, and lower areal mass densities than has been achieved to-date. This could be enabled by ongoing developments in on-orbit manufacturing, whereby large structures are 3D-printed in space from raw feedstock materials. This thesis proposes and analyses a number of attitude control strategies which could be adopted for this next generation of ultra-lightweight, large space structures. Each of the strategies proposed makes use of distributed actuation, which is demonstrated early in the thesis to reduce structural deformations during attitude manoeuvres. All of the proposed strategies are considered to be particularly suitable for structures which are 3d-printed on-orbit, due to the relative simplicity of the actuators and ease with which the actuator placement or construction could be integrated with the on-orbit fabrication of the structure itself. The first strategy proposed is the use of distributed arrays of magnetorquer rods. First, distributed torques are shown to effectively rotate highly flexible structures. This is compared with torques applied to the centre-of-mass of the structure, which cause large surface deformations and can fail to enact a rotation. This is demonstrated using a spring-mass model of a planar structure with embedded actuators. A torque distribution algorithm is then developed to control an individually addressable array of actuators. Attitude control simulations are performed, using the array to control a large space structure, again modelled as a spring-mass system. The attitude control system is demonstrated to effectively detumble a representative 75×75m flexible structure, and perform slew manoeuvres, in the presence of both gravity-gradient torques and a realistic magnetic field model. The development of a Distributed Magnetorquer Demonstration Platform is then presented, a laboratory-scale implementation of the distributed magnetorquer array concept. The platform consists of 48 addressable magnetorquers, arranged with two perpendicular torquers at the nodes of a 5×5 grid. The control algorithms proposed previously in the thesis are implemented and tested on this hardware, demonstrating the practical feasibility of the concept. Results of experiments using a spherical air bearing and Helmholtz cage are presented, demonstrating rest-to-rest slew manoeuvres and detumbling around a single axis using the developed algorithms. The next attitude control strategy presented is the use of embedded current loops, conductive pathways which can be integrated with a spacecraft support structure and used to generate control torques through interaction with the Earth's magnetic field. Length-scaling laws are derived by determining what fraction of a planar spacecraft's mass would need to be allocated to the conductive current loops in order to produce a torque at least as large as the gravity gradient torque. Simulations are then performed of a flexible truss structure, modelled as a spring-mass system, for a range of structural flexibilities and a variety of current loop geometries. Simulations demonstrate rotation of the structure via the electromagnetic force on the current carrying elements, and are also used to characterise the structural deformations caused by the various current loop geometries. An attitude control simulation is then performed, demonstrating a 90◦ slew manoeuvre of a 250×250 m flexible structure through the use of three orthogonal sets of current loops embedded within the spacecraft. The final concept investigated in this thesis is a self-reconfiguring OrigamiSat, where reconfiguration of the proposed OrigamiSat is triggered by changes in the local surface optical properties of an origami structure to harness the solar radiation pressure induced acceleration. OrigamiSats are origami spacecraft with reflective panels which, when flat, operate as a conventional solar sail. Shape reconfiguration, i.e. "folding" of the origami design, allows the OrigamiSat to change operational modes, performing different functions as per mission requirements. For example, a flat OrigamiSat could be reconfigured into the shape of a parabolic reflector, before returning to the flat configuration when required to again operate as a solar sail, providing propellant-free propulsion. Shape reconfiguration or folding of OrigamiSats through the use of surface reflectivity modulation is investigated in this thesis. First, a simplified, folding facet model is used to perform a length-scaling analysis, and then a 2d multibody dynamics simulation is used to demonstrate the principle of solar radiation presure induced folding. A 3d multibody dynamics simulation is then developed and used to demonstrate shape reconfiguration for different origami folding patterns. Here, the attitude dynamics and shape reconfiguration of OrigamiSats are found to be highly coupled, and thus present a challenge from a control perspective. The problem of integrating attitude and shape control of a Miura-fold pattern OrigamiSat through the use of variable reflectivity is then investigated, and a control algorithm developed which uses surface reflectivity modulation of the OrigamiSat facets to enact shape reconfiguration and attitude manoeuvres simultaneously.

Despite the investment in procedure-based simulators in endovascular training programmes, there is a growing concern over the lack of effective haptic feedback, which means that trainees must rely on live practice to develop skills and knowledge of the tactile aspect of Mechanical Thrombectomy. The live procedure carries a significant risk to patients, and the opportunity to carry it out is often limited, which can delay the trainee's development. This research presents a novel tactile controller designed to simulate the physical characteristics of arterial threading for mechanical thrombectomy to assess the efficacy of haptic feedback information generated using force-sensing resistors and game-engine technology for simulating and recording arterial threading. The experimental work was performed in a Virtual Learning Environment to evaluate the effectiveness of the force-sensing box trainer in simulating the physical resistance of an artery in mechanical thrombectomy. The generated force data is presented to a user in a real-time virtual operation theatre indicating the level of applied force. This work demonstrated the successful integration of force-sensor data into a Virtual Learning Environment inside the Unreal Engine. The data can be exported for further analysis by the end-users after completing the simulation to investigate the user's performance. Analysis of the obtained experimental data from the real-time simulation study demonstrated the ability of the system to identify changes to applied force and impacts generated by a user in a real-time simulation using force-sensitive data for simulated arterial threading, indicating the level of applied force exhibited by the user. Supervising practitioners can use the experimental data to determine the number of impacts and force applied during a simulation.

With the drive towards Industry 4.0, the demand for a higher level of automated data capture is increasing. Elevating the quantity and quality of data collected allows for: real time monitoring of processes, purely digital workflows, deeper understanding of processes (particularly, in manufacturing) and stronger links between the digital and physical worlds. In order to achieve the level of data capture required, both an increased quantity of sensors will be required as well as manufacturing methods to automatically integrate these systems in a wide array of increasingly customised components. Additive manufacturing technology lends itself to this application. The body of work enclosed explores methods, strategies, and applications for hybrid additive manufacturing to enable direct write embedding of sensor systems. This can be achieved through a combination of differing materials, or multiple processes to enable materials and manufacturing methods to coexist to create a single component with functionality beyond its structural form. Multi-material fused deposition modelling was identified as a strong candidate for direct writing sensor structures within components, owing to the bottom-up build process and the capability to print conductive composite thermoplastics - explored in Chapter 2 - alongside structural ones. This allowed for a purely digital workflow to be developed, allowing a computer aided engineering model to be manipulated in real time using a physical object with embedded tactile sensors in Chapter 3. In addition, material selection can be tailored to match the contextual requirements of the use case at a localised level. An interdigital dielectric sensor design proposed in Chapter 4 and improved in Chapter 5 through a more refined material selection demonstrates the concept of a localised approach to material selection enabled by hybrid additive manufacturing processes. The sensor enabled real time monitoring of a curing epoxy matrix in the carbon fibre reinforced polymer manufacturing process.

There has been an incredible surge in interest in environmentally friendly technology during the last several decades. Furthermore, wind power is also one of the renewable energy sources that has seen tremendous expansion in recent years, with the installation of new wind farms taking place all over the globe and improvements in wind energy technology that have made this resource considerably more effective. This realisation, along with a greater emphasis on maintaining the long-term sustainability of wind energy systems, has encouraged many to evaluate the prospects wind energy regarding its potential future performance. Floating wind turbines, a recent high-innovation advancement in the wind power industry, are located in deep oceans where fixed-bottom offshore wind turbines are not possible due to foundation costs and a lack of appropriate technology. Besides the design of these new wind turbine towers, the carbon emissions released during their manufacture, installation, operation, and disposal must also be taken into account. Given these new turbine structures, in this study, first the LCA analysis of the barge-type floating wind turbine was focused on and the results were examined. In light of these results, it is aimed to reduce the environmental impact by improving the maintenance and end-of-life steps, which are two important steps, including how to reduce the environmental impact of the turbine. And finally, the cost analysis of the scenarios used in the maintenance and end-of-life steps is made, and both environmental and economic analyses are made and compared. The results of this study indicate that, overall, the barge-type floating wind turbine, which has a 30-year lifetime and utilises an onshore maintenance strategy, has lower environmental and economic impacts than other scenarios. With regard to end-of-life scenarios, the environmental impact of the mechanical recycling process and the cost analysis of incineration scenarios currently seem most appropriate.

Formation of a splash crown and a cavity following the impact of a sphere on a body of liquid is a classical problem. In the related problem of droplet splashing on a flat surface, it has been established that the properties of the surrounding gas can influence the threshold speed at which splashing occurs. At lower impact speeds, this is due mainly to the influence of gas kinetic effects, since the height of the gas lubrication film which is displaced during dynamic wetting is often comparable to the mean free path of the gas. At higher Weber and Reynolds numbers, on the other hand, inertial effects dominate and the density of the gas becomes important in determining whether a splash occurs. Since the influence of the ambient gas on the droplet impact has been known since the beginning of the twentieth century, it is of theoretical importance to investigate whether a similar phenomenon occurs during entry of projectiles into liquids and test the applicability of the wetting theory which is used for droplet impacts. In this work, we present experimental and theoretical investigations of the influence of the ambient gas on cavity formation during a sphere impact on a free liquid surface. In the experiments, a vacuum chamber was used to provide ambient conditions from atmospheric pressure down to 1/100-th of an atmosphere, whist high-speed photography is used to provide detailed images both of the cavity beneath the surface and the splash crown and jets which form above the surface. To elucidate the primary forces at work, we vary the radius of the sphere, the properties of the liquid, the ambient gas pressure, and the type of gas. It is found that the threshold entry speed for cavity formation is influenced by the density of the surrounding gas, whereas changing the mean free path of the gas has no effect. We attribute this phenomenon to the gas slowing the sealing of the thin crown sheet behind the sphere. This assertion is supported with detailed optical measurements of the liquid sheet thickness. In the range of parameters considered, the splash crown influences the movement of the contact line. Finally, we extend our investigations by performing simulations of the impact of a hydrophobic sphere on the surface of a body of water in the presence of ambient air above the liquid. For this purpose, we employ a three-dimensional Eulerian volume of fluid (VOF) multiphase model which uses overlapping meshes. We find that decreasing ambient gas density is not able to suppress cavity formation in the case of hydrophobic spheres.

House prices are a key economic indicator. However, there is usually a delay of two months between the sale of a house and the availability of price data through the Land Registry. Over the past decade, vast quantities of imagery of streets and neighbourhoods have become available online, and the speed at which such imagery is being updated is growing. Simultaneously, advances in deep learning have enhanced our ability to determine the contents of an image automatically. In this thesis, we therefore ask: can we use deep learning to automatically infer house prices from large volumes of photos of houses and neighbourhoods? Focusing on London as a case study, we analyse millions of photos from three sources, covering photographs of streets (Google Street View); crowdsourced photographs of outdoor neighbourhood locations (Geograph); and marketing images of house exteriors and interiors (Zoopla). For each image, we use a convolutional neural network to extract data on the presence of visual features. We investigate whether these features can be used in conjunction with an elastic net regression model to estimate the median house price during 2015 and 2016 for London neighbourhoods. Our results demonstrate that it is possible to automatically infer local house prices from all three sources of imagery. The best estimates are made using Zoopla photographs. This may be due to the larger number of Zoopla images available in our sample (3,727,890 photographs, compared to 519,295 images for Google Street View and 273,999 images for Geograph). However, we find that increasing the number of images available to the model by combining images from different sources has limited benefits. As the speed of collection of such imagery increases, this approach could equip national and local policymakers with more timely information on house prices, enabling them to make more informed decisions about economic policy.

Railway stations are a vital part of the railway system and a critical hub of public transportation. This valuable infrastructure in the heart of cities requires effective risk management processes as well as high-level safety. However, railway stations have faced increasing operational challenges in recent years as a result of the growth in urban areas, high demand, intensive usage, and an increase in train capacity often without any consideration of the size of the existing stations. Consequently, increased risks may exist and lead to train delay, injuries, or fatalities as well as loss with disruption of the whole system and impact on business. Risk management is thus an essential part of the safety system, a system which currently is mostly traditional and does not take advantage of available technologies. The traditional risk process has uncertainties, it takes time, is costly and it depends on experts who are not available all the time. These processes could be developed to cope with future requirements by creating smarter stations and cities, embracing new technologies such as artificial intelligence (AI). It is both wise and necessary to make use of all possible technologies to stop or mitigate the risks and consequences. This can be achieved by improving the risk management process, which will then reflect positively on the quality, reliability, safety, security, as well as cost. This study proposes utilising (AI) methods to improve railway stations' safety and risk management. The methods have powerful analytics and predictive ability. Machine learning (ML) has been proposed using the classification tree, Fuzzy and Nerul network (ANFIS), and deep learning (DL) to identify the risks and people's behaviours in the stations. The information examined is from occurrences at UK stations and ended in casualties. The characteristics of the process and the selected approach are fundamental to selecting the appropriate data. An advanced analytical approach is adopted, using predictions via classification in addition to timely detections. The proposed application presents novel work contributing to the field and aiming to save lives in the future. DL shows a great opportunity to capture many risks in the stations. The outcomes provide a unique contribution where the approach is applicable and makes the process smarter with less human intervention, whilst also supporting decision-makers in real-time. This will also upgrade the risk management to be effective, reflect positively on people's lives and provide a safe environment for all of society.

The microstructure formed after laser powder bed fusion (L-PBF) depends on the thermal gradient and solidification rate with the complex physical interactions of the melt pool and heat flow also impacting the solidifying structure. Due to the layer-by-layer methodology applied in L-PBF, modifying these factors is possible and an area of growing research. Thus, the microstructure itself can be seen as a variable tool to be controlled through processing parameters, allowing L-PBF to be harnessed to move towards the generation of functionally graded microstructures and the formation of components with structures which are fully optimised for their application. This work explores the feasibility of additively manufacturing tailored Inconel 718 microstructures by varying process parameters and scanning strategies, to modify the mechanical properties and performance. This investigation focuses on controlling the heat input and thermal history through process parameter manipulation; notably the heat input parameters (power, scan speed, and hatch spacing) and island scanning parameters (island size, shift, and island overlap). Preliminary investigations into the use of base plate heating at 400°C with the objective of moving from columnar to equiaxed growth were also made. The changes in preferred orientation, morphology and grain size were fully characterised and compared using scanning electron microscopy, electron back scatter diffraction and X-ray diffraction. The solidification cell size was quantified to estimate the effect of the process parameters on the cooling rates. Both grain size and preferred orientation were successfully tailored through scanning parameter modification and separately through base plate heating. Based on these obtained microstructures, graded microstructures were generated using subtle modifications in the island strategy. The viability of using scanning strategies to modify the mechanical properties of Inconel 718 through variations of microstructure was then investigated in more detail. Mechanical testing indicated that the scanning strategy can be used to generate functionally graded specimens; larger grains and more [001] preferred alignment resulting in lower Young's modulus and strength or smaller grains with a reduced preferential [001] orientation along the build direction increased the Young's modulus and strength. Digital image correlation showed that strain in graded structures was concentrated in highly textured areas. The ductility was limited by the entrainment of thin Al oxides folded in the melt pool which were investigated in more detail. Further investigation into the high temperature isothermal oxidation resistance of the microstructurally variant components was made. The mass gain behaviour was found to be broadly similar between the different conditions after 504 h and comparable to wrought Inconel 718 but following sub-parabolic oxidation kinetics possibly attributable to the formation of an underlying Nb rich layer.

It is commonly believed that railway is the safest transportation system for either passengers or freight. Rail operators consistently demand better performance and durability of prestressed concrete sleepers to compete effectively with other transport modes. Track structures guide the safe, economical, and comfortable transit of trains. At present, the use of prestressed concrete sleepers is on the rise around the world as they are more durable, with a life span of 50 years worldwide (70 years in the UK), economical as compared to wood and steel, and easier to manufacture. However, over a long-term service, railway prestressed concrete sleepers are subject to various time-dependent actions, which results in the reduction of durability and capacity. Over decades, the premature failures of traditional railway sleepers have significantly increased track maintenance costs. Due to the need for increased service loads and faster train speed being applied to conventional tracks, the current designs of prestressed concrete sleepers for long-term services may be estimated, largely due to the lack of thorough understanding into the mechanisms of sleeper degradation. In order to study the mechanisms of sleeper degradation over time, significant research effort for railway engineering and technologies has been conducted to perform comprehensive investigations of prestressed concrete sleepers subject to time-dependent actions. The structural performance, material, loading conditions, capacity, support conditions, and environmental conditions of prestressed concrete sleepers have been considered fully. Collaborative research between the University of Birmingham (UoB) and other international railway research institutions such as China Academy of Railway Sciences (CARS), Southwest Jiaotong University (SWJTU), Tsing Hua University (THU), Beijing Jiaotong University (BJJTU) has investigated such important issues as the deformation due to time-dependent actions in prestressed concrete sleepers and capacity of prestressed concrete sleepers. The comprehensive literature review highlights the extremely limited research that has previously been conducted in this field of research. The investigations have been carried out in order to enhance the understanding into time-dependent behaviours of prestressed concrete sleepers. Moreover, this thesis provides an intensive review aimed at predicting time-dependent behaviours in different design codes. Experimental and numerical investigations into the time-dependent behaviours of prestressed concrete sleepers are presented in this thesis. A series of full-scale experiments were conducted in a laboratory and at a railway sleeper factory in order to validate and calibrate the numerical models. The deformation measurements were carried out to predict creep and shrinkage behaviour. Capacity tests has been conducted to identify crack propagation and cumulative damage has been identified in order to calculate fatigue life. The effect of environmental factors on time-dependent behaviours is also presented, while numerical investigations of the time-dependent behaviours of prestressed concrete sleepers have been undertaken to supplement the experimental results in this thesis. The key findings will enhance the development of inspection criteria for railway prestressed concrete sleepers subject to time-dependent actions, improve safety and reliability of railway infrastructures in long-term service.

For large eddy simulation, it is critical to choose the suitable turbulent inlet boundary condition as it significantly affects the calculated flow field. In the thesis, the effect of different inlet boundary conditions including random method (RAND), Lund method and divergence-free synthetic eddies method (DFSEM) on the flow in a channel with a hump are investigated through large-eddy simulation. The simulation results are further compared with experimental data. It has been found that turbulence is nearly fully developed in the case based on Lund method, not fully developed in the case based on DFSEM and not developed in the case based on RAND method. In the flow region before the hump, mean velocity profiles in the case applying Lund method gradually fit the law of the wall as main flow moves towards the hump, but the simulation results based on RAND and DFSEM methods cannot fit the wall function. In the flow region after the hump, cases applying Lund and DFSEM methods could relative precisely predict the size of turbulent bubble and turbulent statistics profiles. While the case based on RAND method cannot capture the positions of flow separation and re-attachment point and overestimates the turbulent bubble size. From this part, it could be found that different turbulent inflow generation methods have a manifested impact on the flow separation and re-attachment after the hump. If the coherent turbulence is maintained in the approach flow, even though turbulent intensity is not large enough, the simulation can still predict the flow separation and turbulent bubble size relative precisely. From the results, even if the simulation based on the Lund method and DFSEM have a better performance than the simulation based on the RAND method, the results still cannot agree well with the experimental data. There are some possible reasons that result in the big difference. Firstly, the spanwise width of the simulation domain is relatively small. Additionally, the LES sub-grid scale model has a slight impact on the results according to the previous research. Finally, the DFSEM is sensitive to the surface normal gradient schemes. Beside the choice of the turbulent inlet boundary condition (IBC) methods, the settings of each turbulent IBC method are critical as well. In this thesis, the effect of setting different IBC methods in LES using the Lund method, the divergence-free synthetic eddies method (DFSEM) and the digital filter method (DFM) on the simulation of the boundary layer over a flat plate is investigated. This research also fully explained the influences of different IBC methods on the results of turbulent kinetic energy budget terms, and it is found that the DFM and the DFSEM both have good performance. In addition, for the DFM and DFSEM, the input parameter such as turbulent length scales are hard to set generally without prior knowledge of the flow field. It is found that the simulation results based on these two methods with constant turbulent length scales of 0.4-1.0 times of the value of boundary layer thickness could agree well with the DNS results to a great extent after about 10 boundary layer thickness along the streamwise direction. Overall, it could be recommended that the DFM and the DFSEM can be used in this case with the constant input turbulent length scale about 0.4-1.0 times of the value of boundary thickness.

Leveraging blockchain technology, IoT data can be recorded as immutable transactions and processed in consensus by blockchain nodes. Blockchain can ensure distributed and secure IoT data management due to its inherent features, such as transparency, auditability, traceability, and accountability. However, the implementation of blockchain in IoT systems is still facing some challenges. First, IoT systems are data-driven, characterized by high velocity, high volume of data, and high mobility, making data security an issue. Next, blockchain presents technical constraints of a complex nature, such as limited space, immutability, and excessive computational power, that can limit its adoption in IoT systems at scale. Therefore, to address these challenges, a comprehensive investigation of architectural knowledge, design decisions, architectural tactics, styles, and data allocation mechanism that can drive the architectural design of IoT systems supported by blockchain is required. In this work, we identify the common quality attribute requirements, design decisions, and tradeoffs and their impact on system goals. We also present a catalog of architectural tactics that can help architects in achieving the quality attribute requirements of the system. In addition, we codify a set of reference architecture styles and variants for IoT systems supported by blockchain. Using a case study of healthcare, we evaluate the general fitness of styles with respect to quality attribute requirements using the Architecture Tradeoff Analysis Method (ATAM) and simulation. Finally, we propose a data allocation mechanism to dynamically decide on on-chain and off-chain data storage. The significance of this study is that it informs architects and designers with guidelines and blueprints on the architectural design of this category of systems by introducing a systematic investigation and evaluation approach.

The advent of computer simulation has not only reduced the cost and time for testing a suitable design or product through expensive experiments but also become a potential tool for an in-silico and in-vitro research. With computer simulations different solutions can be evaluated numerically and only promising designs are sent to laboratory or production line for scale-up production. In this thesis, flow problems related to cardiovascular disorders in humans were studied using a mesh-free particle modelling technique. Since cardiovascular disease is the main cause of death worldwide, study into its causes, progression, and treatment is necessary. Therefore, the diseased cardiovascular organs (aortic valve and coronary artery) which are solid, and the blood (liquid) are represented as particles, constituting the full geometry that is employed for this study. This takes us to the use of a framework known as Discrete Multiphysics, which integrates two particle approaches, notably the Lattice Spring Model for solid mechanics and Smooth Particle Hydrodynamics for fluid mechanics. Employing this approach therefore, the mesh-free particle method in the form of discrete multiphysics was used to simulate the two cardiovascular problems with complex geometries, (i) effect of aortic valve calcification on blood flow and deformation of the valve (ii) the interaction between fluid (blood) and solid (arterial wall and stent) within a coronary artery implanted with a metal stent. In the case of aortic valve, the effects of various calcification stages (ranging from mild to severe) on cardiac output were simulated and evaluated. It was found that increasing levels of calcification result in a reduction in heart flow rate, and a critical threshold of calcification exists below which the flow rate drastically declines. Additionally, in the case of coronary stents, the key factors that affect how much the stent deforms are sorted according to dimensionless numbers, and a connection between the elastic forces of the stent and the pressure forces of the fluid is formed. The blood flow and stiffness of the stent material were subsequently found to considerably contribute to the stent's deformation and to have an impact on its rate of deformation. In both cases however, mechanical stresses were analysed and how the severity of calcification as well as the elastic and pressure forces affect the rate of deformation of the valve and the stent respectively were investigated. Therefore, it will be convincingly concluded that despite the complexity of these biological organ's geometries, we were able to achieve the solid-liquid mechanics analyses with little difficulty, thanks to Discrete Multiphysics modelling approach.

Proton Transfer Reaction-Mass Spectrometry (PTR-MS) has been in use since the 1990s for real-time measurement of trace volatile organic compounds (VOCs) from parts per million by volume down to sub-parts per billion by volume in a wide range of applications such as environmental monitoring, health, homeland security and food sciences. The work presented in this thesis deals with new developments and applications of PTR-MS and has different interconnected themes. The first is hardware development; principally the testing and characterisation of new radio frequency (RF) ion-funnels within the PTR reactor designed to improve ion transmission and thus sensitivity of detection for trace compounds. The desire to characterise better the effect of RF fields on analyte ions in the reactor led to a fundamental study in which the proportion of analyte product ions was determined as a function of the reduced electric field strength, first without the application of RF fields. This was to compare their behaviours when subject to RF fields. An unexpected and previously unreported shift in these product ion distributions was discovered, depending upon how the reduced electric field, E/n, was created. A second theme is the creation of a methodology for, and production of, a mass spectral library for individual VOCs analysed by PTR-MS. This was to assist in the development and testing of a new software algorithm for compound identification and quantification of complex VOC mixtures when no pre-separation technique such as gas chromatography (GC) is used. The third theme is the application of both these hardware and software developments in the practical analysis of VOCs, not just in ambient air but also in water, where a portable water sampler has also been developed, so that impurities in water could be determined with high specificity and sensitivity in real-time. This has been used in tandem not only with PTR-MS, but also with two more conventional portable electron impact mass spectrometers (EI-MS). One of these analysers was taken to Mainz, Germany, as part of a secondment, to investigate the possibility of using the instrument instead of the more expensive and complicated PTR-MS for determination of a parameter of interest to atmospheric chemists: 'the total OH reactivity' of the atmosphere. The work in this thesis was conducted as part of a Marie Skłodowska-Curie Actions Innovative Training Network (ITN): Ion-Molecule Processes for Analytical Chemical Technologies (IMPACT). This European ITN programme was named through the European Commission's HORIZON 2020 Programme. The work was directly influenced by the desire to improve analytical tools available to a commercial analytical instrument manufacturer, Kore Technology Ltd, at whose premises most of the work took place.

Both mechanical and chemical finishing process are regularly used to reduce frictional losses and improve the wear resistance of components within mechanical systems like internal combustion engine cylinders and fluid pumps. Coatings often require highly specialised equipment and are usually expensive in comparison to mechanical processes. In many applications, both mechanical finishing and chemical coatings are applied to component surfaces in order to achieve a surface with all the desired properties. This thesis applies tribofilm forming research when using particular nano-additives to mechanical finishing processes in order to mechanochemically generate a tribologically superior surface through a simple, single stage operation. Cast iron plates were honed against grit paper with dry Cu₂S microparticles causing a reduction reaction in the particles for sulphur to then react with the iron in the plates. Energy Dispersive Spectroscopy (EDS) revealed a huge increase in the quantity of copper in the cast iron caused by the mechanochemical honing. X-ray Photoelectron Spectroscopy (XPS) did suggest a small quantity of iron sulphides had formed. While the honing process did successfully reduce the copper sulphide, a lot of the sulphur was lost in the process while the copper sintered into the surface. Lubricated friction tests at a variety loads showed a small reduction in friction with the copper honed samples. In order to better understand mechanisms for tribofilm formation and discover any synergistic behaviours between nano-MoS₂ and ceramic nanospheres, SiO₂ was tested with MoS₂ as lubricant additives in a variety of High Frequency Reciprocating Rig (HFRR) tests. At high load and 50 °C the two nanoparticles on their own could not improve friction or wear but when used together, the SiO₂ improved the conditions for MoS₂ to chemically react with the steel contacts forming a protective tribofilm. This was characterised through a combination of EDS and Hard X-Ray Photoelectron Spectroscopy (HAXPES). Attempts to recreate these results in the TE 77 High Frequency Friction Machine (TE 77) low load adapter were unsuccessful. Keeping lubricant additives successfully dispersed is a well documented issue. To get round this, Cu₂S and Al₂O₃ microparticles were mixed with a base grease whose viscosity was great enough that once dispersed the particles could not settle. The formulated greases were put under very high load (920N) in line contact in the TE 77 to generate a tribofilm style layer that was then tested with point contact wear in a base lubricant. The formulated greases were tested against a commercial grease containing MoS₂. At low loads, only the MoS₂ treated surface produced lower friction than the untreated plate. In 50 °C tests at a higher load, all the treated surfaces slightly reduced friction compared to the untreated plate and the hybrid Cu₂S, Al₂O₃ surface had the lowest friction. MoS₂ produced the shallowest scar with a tribofilm similar to that generated with the hybrid lubricant in the previous chapter. This thesis introduces a proof of concept for two mechanochemical finishing processes as well as developing understanding for the synergistic behaviour of Transition Metal Dichalcogenides (TMDC) and ceramics as nano-additives in lubricants. The potential scope for this research is huge as the finishing processes could be beneficial to any mechanical system where there are sliding metal contacts. A large amount of work is still needed to bring this work from its concept stage to full realisation in a commercial setting.

In recent years, the concept of vehicle platooning has gained widespread attention for its highly efficient road usage and lower fuel consumption. However, the aerodynamics of vehicle platoons travelling in a tunnel are not well understood, even though more and more road tunnels have been built to alleviate the traffic congestion problem. This research aims to improve our understanding of the aerodynamic phenomena associated with a long vehicle platoon running through a tunnel. The effect of the tunnel existence, blockage ratio, symmetry of the traffic lane and the inter-vehicle spacing on the aerodynamics performance of the long platoon will be investigated. To achieve this goal, both model-scale experiments and numerical simulations (IDDES) were conducted, and the results are compared to a similar study conducted in the open air. The slipstream velocity and pressure, the lorry surface pressure, as well as the drag coefficient, were investigated systematically. The results show greater pressure variations when the platoon is running through the tunnel than in open air. The piston effect in the tunnel leads to a lower approaching velocity and a weaker flow separation compared to the case in the open air. All vehicles, in both the tunnel and the open air, experience a drag reduction due to platooning. Interestingly, the drag reduction in the tunnel is 20% greater than that in the open air, implying a greater potential in fuel saving. When the blockage ratio decreases, the piston effect becomes less effective, making the flow field and the variation drag reduction ratio approach the pattern observed in the open air. Unlike a single lorry, increasing the blockage ratio does not lead to an increase of the drag of every lorry in the platoon. In a small tunnel, the some intermediate lorries have smaller drag coefficients, while others have larger drag coefficients compared to the lorries in large tunnels. Therefore, the overall drag coefficients of the platoons are not affected by the blockage ratio, except for the platoon with 0.1L spacing. It is further found that the travelling on the asymmetrical traffic lane may results in some asymmetry of the flow field, but has little influence on the drag and side forces experienced by the lorries. In the open air, the drag coefficients of all lorries are monotonically decreasing with the spacing due to the stronger shielding effect. When the spacing reduces to 0.25L, the intermediate lorries have larger drag coefficients than the same lorries in the platoon with larger spacings due to the change of the wake structure and the confinement of tunnel walls. Therefore, the overall drag reduction ratio does not monotonically increase with the decreasing spacing. The inter-vehicle spacing control strategy should be reconsidered when a long platoon travelling from the open air into a road tunnel.

Ride-sharing is a promising solution for transportation issues such as traffic congestion and parking land use, which are brought about by the extensive usage of private vehicles. In the near future, large-scale Shared Autonomous Mobility-on-Demand (SAMoD) systems are expected to be deployed with the realization of self-driving vehicles. It has the potential to encourage a car-free lifestyle and create a new urban mobility mode where ride-sharing is widely adopted among people. This thesis addresses the problem of improving the efficiency and quality of vehicle dispatch in high-capacity SAMoD systems. The first part of the thesis develops a dispatcher which can efficiently explore the complete candidate match space and produce the optimal assignment policy when only deterministic information is concerned. It uses an incremental search method that can quickly prune out infeasible candidates to reduce the search space. It also has an iterative re-optimization strategy to dynamically alter the assignment policy to take into account both previous and newly revealed requests. Case studies of New York City using real-world data shows that it outperforms the state-of-the-art in terms of service rate and system scalability. The dispatcher developed in this part can serve as a foundation for the next two parts, which consider two kinds of uncertain information, stochastic travel times and the dynamic distribution of requests in the long-term future, respectively. The second part of the thesis describes a framework which makes use of stochastic travel time models to optimize the reliability of vehicle dispatch. It employs a candidate match search method to generate a candidate pool, uses a set of preprocessed shortest path tables to score the candidates and provides an assignment policy that maximizes the overall score. Two different dispatch objectives are discussed: the on-time arrival probabilities of requests and the profit of the platform. Experimental studies show that higher service rates, reliability and profits can be achieved by considering travel time uncertainty. The third part of the thesis presents a deep reinforcement learning based approach to optimize assignment polices in a more far-sighted way. It models the vehicle dispatch problem as a Markov Decision Process (MDP) and uses a policy evaluation method to learn a value function from the historic movements of drivers. The learned value function is employed to score candidate matches to guide a dispatcher optimizing long-term objective, and will be continually updated online to capture the real-time dynamics of the system. It is shown by experiments that the value function helps the dispatcher to yield higher service rates.

The quantification of key microstructural parameters as a function of creep conditions is commonplace in the assessment of Grade 91 and other creep strength enhanced ferritic (CSEF) power-plant steels. This thesis analyses a broad range of Grade 91 steels which include, ex-service, developmental and more recently produced steels; each distinct in terms of creep performance and composition. Characterisation was performed using a combined series of techniques that provided through-length scale chemical analysis and site-specific quantification. Initial steps involved quantitatively assessing the chemical homogeneity of the steels using micro-X-ray fluorescence (μ-XRF) and scanning electron microscope (SEM) based chemical mapping. All steels displayed micro-segregation although it was greatest in the ex-service steels. In these steels the micro-hardness was found to vary from 190 to 225 HV primarily due to compositional heterogeneity caused by micro-segregation. Focussed ion beam (FIB-SEM) imaging and thin foil lift-outs analysed using scanning transmission electron microscopy (STEM) chemical mapping showed area coverages of M23C6 carbides and V-rich carbo-nitrides were greater in positively segregated regions. Site-specific quantification using SEM-back scattered electron (BSE) imaging of the Laves phase showed a difference of up to a factor of two in the number density of particles and area coverage due to micro-segregation. For ex-service steels this was consistent with thermo-dynamic predictions. A procedure for quantification of inclusions was developed which considered optimal voltage parameters investigated through studies and simulations, micro-segregation effects and fine (down to 0.25 μm) inclusions. These are all important deviations from conventional automated SEM procedures. This work shows that micro-segregation can have significant effects on micro-hardness and characteristics of various second phase particles. Since the performance of a material is likely to be controlled by local microstructural extremes rather than average properties, these findings have important implications for the characterisation and failure assessment of Grade 91 and other CSEF steels.

Nickel superalloys find several applications in the aerospace and automotive industry, notably as safety-critical components such as turbine blades and turbine discs. A new kind of manufacturing technique termed Additive Manufacturing (AM) has gained significant interest due to its capabilities of fabricating components that could minimise material wastage and/or components that possess complex geometries, possibly unmanufacturable using conventional technologies. Among the various AM processes for metallic components, powder bed fusion (PBF) has shown great potential for manufacturing Inconel 718 (IN718) components. The future of PBF is undoubtedly promising but there are still some gaps to fill before AM components materials are ready for widespread industrial application. Laser shock peening (LSP) is an effective surface enhancement process that is used in an aeroengine manufacturing line to strengthen components resistance to surface-related failures. Pioneering work in the late 1990s employed synchrotron radiation for residual stress assessment to enable the adoption of LSP in aeroengine components. The technology is now used in engines worth tens of billions of pounds. Back to the present time, there is a push to adopt AM technology as a step-change in production and integrating LSP into AM IN718 seems to be an obvious way forward. The work presented in this thesis aims to better understand the effects that LSP has brought to AM IN718, which have not been systematically investigated. Microstructures of AM IN718 have been characterised with scanning electron microscopy (SEM) using the electron backscattered diffraction (EBSD) technique and transmission electron microscopy (TEM) to correlate microstructural changes with the process conditions. Mechanical properties (e.g. microhardness and tensile properties) have also been measured and rationalised with regards to the LSP treatments. Specimens were put under various thermal and mechanical loading to simulate the aeroengine setting. Synchrotron X-ray diffraction (SXRD) was utilised to assess the residual stress relaxation of these specimens and the phases present in AM IN718. Full width half maximum (FWHM) data from the SXRD experiment were used to characterise the cold work present in the specimens that were LSP-ed. Three point flexure test was selected as the appropriate test scenario to evaluate the mechanical performance of the LSP AM IN718 at a structural level. Comparison of the observed mechanical performance indicates that there is potential for LSP to be integrated with the AM process of IN718. Within a 550°C environment, the LSP-ed specimens hold up to 50% of its initial residual stress after going through about a million flexural cycles. Surface roughness and cold work were found to be key factors controlling the fatigue performance. Microstructure analysis suggests that LSP bring minimal cold work to AM IN718 which is the contributing factor to it's excellent containment of residual stress. Results from this licentiate thesis provide the basis for the further industrial research on the cyclic mechanical properties of AM IN718.

A range of metal-organic frameworks (MOFs) have been synthesised and characterised by a variety of techniques, primarily including powder X-ray diffraction (PXRD), thermogravimetric analysis (TGA), UV-Vis and Scanning Electron Microscopy (SEM). Some of these MOFs have been investigated for applications related to CO2 capture and catalysis. A mixed-metal MOF, MUV-10, comprised of titanium in combination with Ca or Mn, is thoroughly investigated with the aim of making modifications, including the substitution of some Ti for Sn. Synchrotron X-ray absorption spectroscopy confirmed the successful incorporation of Sn(IV) into the mixed-metal MOFs. Gas adsorption studies were also completed and a brief investigation into the modification of the organic linker used in the material, with effects on the band gap of the material discussed. The modified MUV-10 MOFs have been extensively analysed and their activity tested in the photoreduction of carbon dioxide, the reduction of carbon dioxide to cyclic carbonates using epoxides and in the conversion of glucose to 5-hydroxymethylfurfural (5-HMF). For the formation of cyclic carbonates using epoxides and carbon dioxide, MUV-10 offered excellent conversion, recyclability, and stability. For the conversion of glucose, when small quantities of HCl were used, MUV-10 demonstrated a greater production of 5-HMF compared to the UiO-66 reference material, with enhanced performance for MUV-10 substituted with Sn. The synthesis of novel Ti MOFs has been attempted, and although none were produced in a form for structure solution, some promising materials were identified for future study. Similar studies on the synthesis of Sn(IV) MOFs, not yet known in the literature, also proved not to form large enough crystals for structure solution, although a new anionic Sn(II) coordination polymer based on the chelidamate ligand was crystallised.

Intelligent transportation systems have the potential to provide road users with a range of useful applications, including vehicle preconditioning, traffic flow management and intelligent parking recommendations. The majority of these applications can benefit from knowledge of vehicle activities (common situations that a vehicle encounters e.g. traffic), along with the upcoming destinations that a vehicle will visit. We focus on the trajectories that vehicles provide, and the data contained within them, in order to ascertain information about the patterns in individuals' mobility data. Machine learning has been used in many different vehicle applications, and we focus on using these techniques to predict the activity of a vehicle and its future destinations. Clustering methods can be applied at the level of trajectories or the individual instances within them, and we explore both of these alternatives in this thesis. Additionally, we explore several classification approaches to predict activities and destinations. In developing our methods, we make use of a combination of both geospatial and temporal data along with on-board vehicle sensor data. This thesis presents novel methods for filtering stay points to identify points of interest and applying destination prediction to vehicle trajectories. Existing methods for stay point detection are not specific to vehicles, and therefore any region of low mobility is potentially considered to be of interest. We propose a novel method for filtering the extracted stay points to identify points of interest, using vehicle data to predict vehicle activities. The predicted activities are further used to represent trajectories as sequences of annotated locations, to inform the detection of similarities between journeys. Finally, this thesis presents a novel method for using additional properties of a trajectory to cluster trajectories into groupings of similar trajectories with the aim of improving the accuracy of destination prediction. We evaluate our proposed methods on a set of vehicle datasets, varying in purpose and the data available.

Next-generation wearable devices need lithium-ion batteries (LIBs) designs that are mechanically flexible, safe and provide stable energy sources. In conventional battery designs the electrodes are susceptible to mechanical damage when exposed to cyclic flexure leading to capacity loss, resistance increases and severely limited cycle life. Typical electrode binders use polyvinylidene fluoride (PVDF) which is brittle and has low adhesion thus unsuitable for flexible LIBs. The use of polyurethane (PU) as a potential cathode binder was explored because of its promising mechanical properties for flexible LIBs. Three potential PU chemistries were investigated, and ether-based PU was selected due to minimal swelling in the electrolytes thus retaining its mechanical properties. It further demonstrated stable electrochemical cycling (>90% capacity retention after 250 cycles at 1C discharge rate), low electrical resistance, high toughness (>200 J cm−3 compared to 1.7 J cm−3 for PVDF) and high elasticity making it a promising polymer choice for flexible electrodes. Conventional metal foil current collectors were found to have poor elasticity, but carbon nanotube (CNT) fabrics demonstrated the potential to make mechanically resilient electrodes. Current collectors using CNT fabrics were integrated with an optimised ether-based PU electrode binder and these cells were investigated. CNT fabrics were found to provide enhanced electrode/current collector interface properties that enabled higher rate performance (80% capacity retention at 5C discharge rate) than Al foil electrodes and improved cohesion and structural integrity of the electrode composite. An absence of standardised methods for testing the cyclic flexure performance of flexible LIBs led to the design and manufacture of a bespoke fatigue flexure rig and the development of new testing methods. Commercially available flexible cells and the novel electrodes described above were then compared. The electrodes of the commercial cells suffered cracking and material detachment during flexure cycling, with complete fatigue failure of the metal foil current collectors. Whereas, the developed CNT fabric and ether-based PU electrodes demonstrated no capacity loss after 3000 flexure cycles and minimal resistance increase with the CNT fabric electrodes yielding a significantly higher energy density than commercial cells.